APPENDIX 5. ORNITHOLOGY.

Appendix 5.1. Earlier ornithological survey work in the Uluguru Mountains.

The reader of this account should bear in mind that very few of the earlier papers give any clue about the intensity of the research at study sites, and much uncertainty exists about the exact study sites in the older literature. We cannot guarantee that the list of earlier activities in the appendix is complete but hope that the account serves to give an overview.

The first ornithological collections date from the 1890’s. Specimens from the eastern part of the Ulugurus, including Mhonda, were collected by Stuhlman and delivered to the Museum of Berlin. No report was ever compiled from his collections, but 14 species were mentioned in various publications by Reichenow from 1889, 1894, 1895 and 1900-1905 (Friedmann and Stager 1964). In 1913 Ludwig Schuster collected specimens in Bunduki, Nyandiduma and Mamba in the western part of the Ulugurus and in Mseru in the eastern part, but only 14 species were published (Schuster 1926, Friedmann and Stager 1964). Arthur Loveridge visited the Ulugurus at least as early as in 1918 but the only published note from this year is a description of a Paradise Flycatcher nest (Loveridge 1922). In June 1920 C.M.F. Swynnerton collected 69 specimens of 37 species, which were delivered, to the Natural History Museum in Tring, England (Friedmann and Stager 1964). Friedmann and Loveridge (1937) mention that Swynnerton had a local collector.

Serious ornithological exploration of the Uluguru Mountains really began as recently as 1921 and 1922. In May-June 1921 and May-July 1922 Salimu bin Asmani, a trained native collector employed by Arthur Loveridge, made the first sizable Uluguru collections (98 species, of which c. 40 are forest species), chiefly around Bagilo (Friedmann and Stager 1964), which was his home village, but also in other parts of the Ulugurus, including Mbeta (Friedmann and Loveridge 1937, Friedmann and Stager 1964). The major part of this collection is now in the Museum of Comparative Zoology, Harvard University, USA, and was reported on by Friedmann and Loveridge (1937).

In 1926, Loveridge and his (bird) collector Salimu collected 77 species of birds (c. 45 forest species) during a survey where the main emphasis was put on reptiles and amphibians (Friedmann 1928, Friedmann and Stager 1964). One of the species was unknown to science: the Uluguru Bush Shrike Malaconotus alius. All specimens later came to the Museum of Comparative Zoology, Harvard University, USA, and were reported on by Friedmann (1927, 1928 and 1929). The specimens were labelled Bagilo (11-28 September), Nyange (1-11 October), Nyingwe[1] (1 October, 8 October, 13-19 October), Tawa (8 October), Mkaraji (20-22 October) and Vituri (26-29 October) (Friedmann 1928).

In 1937 R. E. Moreau spent some time in Kibungo Forest (a bit north of or synonymous with Kimboza Forest) and Kinole Forest (Moreau 1938, Friedmann and Stager 1964), resulting in the discovery at Kinole of a species unknown to science: Mrs Moreau's Warbler Scepomycter winifredae (named in honour of his wife Mrs Winifred Moreau). Stuart and Jensen (1985) mention specimens collected by Moreau in the Ulugurus "mainly in 1938", most of which are now in the Natural History Museum, Tring, England. Moreau also visited the western part of the Uluguru South F.R. for some few days and later sent his collector Charles Abdallah to the same locality to obtain specimens of an unknown species he had heard there (Moreau 1946). Mr. Abdallah drew the conclusion that the unidentified voice belonged to Mrs Moreau's Warbler Scepomycter winifredae.

N.R. Fuggles-Couchman visited many parts of Tanzania over many years in connection with his job as agricultural officer (Fuggles-Couchmann 1939, 1984a,b,c; obituary in Ibis 136 p. 109). His selected notes mention a few observations from Mgeta, Kasanga, Mngazi and the eastern foothills (including Kibungo Forest) (Fuggles-Couchmann 1939 and 1984a,b,c; Stuart and Jensen 1985).

During twenty years of residence on Tanzanian sisal estates from 1947 to 1967, Thorkild Andersen from Denmark collected skins at various localities (in total at least 11685 specimens of 690 species, Britton 1981). He lived in Morogoro for some years and visited the Ulugurus several times, including Kimboza. Preliminary research on Andersen's work has revealed that he had two local collectors with him at the estate in Morogoro (J. Fjeldså pers. comm.). Friedmann and Stager (1964) mention a list Andersen sent them of birds obtained for him in the Ulugurus by his native collectors.

Britton (1978) briefly reviewed 3005 Andersen specimens collected in Tanzania and now stored in museums in Leiden (1749), Stuttgart (971), Basel (280) and New York (5). The ones mentioned from the Ulugurus were labelled: Bagilo (1800 m), Tegetero (900 m) and Uluguru Mountains (mainly 800-1000 m). Britton (1981) later reported on some Andersen skins stored in Munich (3200) and Bonn (1578), including some from the Ulugurus. He did not, however, report on the skins stored in museums in Copenhagen (3949) although he examined the collection. This was due to an agreement with N. E. Franzmann of ZMUC who already planned to publish on the specimens. Unfortunately, Franzmann passed away before writing anything, but the specimens in Copenhagen will be published within the next few years (Fjeldså and Dinesen in prep.). A sporadic study of the collection in Copenhagen in connection with writing this report revealed that the majority of the Uluguru specimens were labelled Bagilo 1800 m or simply Uluguru Mountains. There were, however, also a few specimens from the Mgeta area in the western part of the Ulugurus.

The high number of Bagilo specimens indicate that Andersen or his collectors spent a good deal of time in this area which is on the eastern slopes of the Uluguru North F.R north of our study area at Tegetero (see Figure 1.5). For some of the forest species labelled Bagilo 1800 m this seems an unusually high elevation. Furthermore, some non-forest species, e.g. Waxbill Estrilda astrild, Yellow-bellied Waxbill Estrilda quartinia and Cape Robin-Chat Cossypha caffra, are labelled Bagilo 1800 m (Stuart and Jensen 1985) though this area is deep inside montane forest according to our 1:50.000 map, which is based on air photographies from 1964. It is therefore likely that his specimens labelled 1800 m have been collected over a wide elevational range. There may, however, have been a camp inside the forest at 1800 m altitude. In this connection it is interesting that also the Bagilo specimens mentioned in Loveridge (1960) were labelled 1800 m. The Andersen specimens labelled Uluguru Mountains could be from anywhere in the Ulugurus, since Andersen did not write diaries (J. Fjeldså pers. comm.) and never published anything himself.

Among the most interesting Andersen specimens are at least seven specimens of Malaconotus alius collected between 1952 and 1961 (six labelled Bagilo; Britton 1981), five specimens of Anthreptes rubritorques (the only proofs of its presence in the Ulugurus; Britton 1978 and 1981, Stuart and Jensen 1985), one specimen of Ploceus nicolli (Franzmann 1983, Stuart and Jensen 1985) and (Copenhagen specimens only) seven specimens of Scepomycter winifredae. As a curiosity we can mention that Mze Leonard Bansi from Tegetero, whom we employed as a guide on our work west of Tegetero in December 1993, worked with Thorkild Andersen in 1948-49 when 23 years of age.

J.G. Williams, curator of the bird collection of the Coryndon Museum in Nairobi, Kenya, visited the Uluguru Mountains briefly in October 1948 and November 1950 (Williams 1951). Stuart and Jensen (1985) mention visits by Williams between 1948 and 1951. Most of Williams' specimens are now in the National Museum of Kenya, Nairobi, and the Natural History Museum, Tring, England (Stuart and Jensen 1985). One of the localities he visited was Bunduki.

From 22 November 1961 to 15 February 1962 Gerd H. Heinrich collected specimens for the Peabody Museum of Natural History, Yale University, in the lowlands of the Uluguru Mountains around Morogoro and in cloud forests at 1500-2000 m (Ripley and Heinrich 1966, 1969). Ripley and Heinrich (1966 and 1969) later reported on 16 of the species collected, including 13 forest species.

From 6 to 15 January 1964 the Cheney expedition collected specimens (56 species, about half of them forest species) and investigated the area at Bunduki at an elevation of about 1530 m (Friedman and Stager 1964).

January 1980 S. N. Stuart visited Morningside and the northern slopes of Lupanga in the Uluguru North F.R. (Stuart and Jensen 1985). In June-July 1981 Scharff et al. (1982) studied the invertebrate fauna (N. Scharff and M. Stoltze) and the avifauna (F.P. Jensen and S.N. Stuart) of the Uluguru Mountains. The following sites were visited by Stuart and Jensen:

· Uluguru North F.R.: The western side of Lupanga from the forest edge at 1400 m to the peak at 2138 m and down to 1200 m on the eastern side in July.

· Kimboza Forest (lowland forest at 300 m altitude). A three-day visit in July.

· Uluguru South F.R.: The forest above the village Tchenzema was investigated and excursions to the western scarp of the Lukwangule Plateau were made (1900-2400 m) in July.

In September 1982 F. P. Jensen made a brief visit to Lupanga Peak and the eastern slopes of Lupanga down to 1300 m and in November 1984 he made a brief visit to Kimboza Forest (Stuart and Jensen 1985).

Observations from later brief visits to the Ulugurus by N. E. Baker, K. Howell, D. C. Moyer, R. J. Stjernstedt and D. A. Turner were included in Stuart and Jensen's paper from 1985. Their visits include Bunduki and Morningside. We have made no attempt to find out whether the Ulugurus have been visited since the publication of Stuart and Jensen's paper from 1985. In this connection it should be mentioned that from 1984 and until recently it has been difficulty for foreigners to gain permit to work in the Ulugurus.

The Ulugurus were included in biogeographical analyses carried out by Stuart (1981a and 1983) and by Stuart et al. (1993).

Appendix 5.2. Footnotes and definitions, Table 5.1.

Unpublished information from J. Fjeldså was used to compile the species list for the Rubeho Mountains and to a lesser extent the Pare Mountains. For other ranges we have used various publications.

Key, Tanganyika-Nyasa Montane Forests: Taita: Taita Hills and Mount Kasigau, Southern Kenya. Pares: Pare Mountains, Northern Tanzania. Usambaras: Usambara Mountains, Tanzania. Ngurus: Nguru Mountains, Tanzania. Ukagurus: Ukaguru Mountains, Tanzania. Ulugurus: Uluguru Mountains, Tanzania. Udzungwas: Udzungwa Mountains, Tanzania. S. Highl.: Southern Highlands of Tanzania. N. Malawi: Northern Malawi (includes mountain ranges north of 14°S, e.g. Misuku Hills, Nyika Plateau [a part of Nyika Plateau is in extreme NE Zambia], and Viphya Mountains). S. Malawi: Southern Malawi (includes mountain ranges south of 14°S, e.g. Shire Highlands, Mount Mulanje and Mount Thyolo). Mt. Namuli: Mount Namuli, Mozambique. Mt. Chiper.: Mount Chiperone, Mozambique. Njesi Pl.: Njesi Plateau, Mozambique. +: Forests outside the Tanganyika-Nyasa mountain range, these are explained under “Key, footnotes”.

Certain areas, e.g. Northern Malawi, could be split into more biogeographical units, but the aim of Table 5.1 is only to give a rough overview. Especially the montane ranges in Mozambique are poorly known. See also (2) below.

Key, categories of threat: Threatened (Th): With category Threatened in Collar et al. (1994). (The category Threatened is further subdivided in Collar et al. (1994). For simplicity and because the book reached us a little late in the writing phase these subdivisions are not included here. Near-threatened (Nt): With category Near-threatened in Collar et al. (1994). Restricted-range but not Threatened or Near-threatened (xx): Not considered Threatened or Near-threatened in Collar et al. (1994) but with an estimated breeding range of less than 50,000 km². The estimations of breeding range are our own, see (1).

Key, footnotes given in Table 5.1: 1: Also in the Mahenge Highlands, isolated south of the Udzungwas and belonging to the Tanganyika-Nyasa Montane Forest Group, and in coastal Mozambique, near Beira. 2: Also in the Matengo Highlands, SW Tanzania, which are situated immediately south of the Southern Highlands. 3: Sokoke-Arabuko forest at the coast of Kenya (main population). 4: Three forests in eastern Zimbabwe (Chirinda Forest, Vumba Highlands, Stapleton) and Mount Gorongoza, Mozambique. 5: Coastal Sokoke-Arabuko Forest, Kenya (main population). 6: Certain montane forests in Kenya. 7: Several coastal forests of Tanzania, Mozambique and Kenya. 8: These are not given full species status in Dowsett and Dowsett-Lemaire (1993). 9: Earlier records exist from gallery forest around the lower Tana River, Kenya. It may still survive here in remaining forest patches though sometimes considered “possibly extinct”. 10: This remarkably elusive and little known migratory thrush also occurs (at least seasonally) in certain coastal forests of Kenya, Tanzania, northern Mozambique and South Africa. Two races are known from single specimens from Zaire and Sudan. 11: Discovered as recently as in 1991 (Dinesen et al. 1994). 12: Several coastal forests in Kenya, Tanzania (including Zanzibar) and Mozambique. 13: Coastal woodlands and forests from Southern Somalia south to Natal and Zululand in South Africa; inland in Voi, Kenya and Ruaha N.P., Tanzania. 14: Forests and well-wooded country in coastal lowlands from the lower Jubba Valley, Somalia, to Wami River, Tanzania (with records from Zanzibar in the 1930ies). Inland in Shimba Hills, Kenya, and on Mafi Mountain, Tanzania. 15: Also in the Uvidundas (south of the Rubehos). 16: Forests (locally moist bush land) in coastal Kenya and Tanzania (thinly distributed); also Shimba Hills, Kenya.

1. A “restricted-range species” is defined as “a species with an estimated worldwide breeding range of less than 50,000 km² “. The biodiversity mapping project described in ICBP (1992) was based on an assessment of the occurrence of restricted-range species, a coming BirdLife publication (The BirdLife Directory of Endemic Bird Areas, in prep.) will also be based on restricted-range species. Almost all Threatened and Near-threatened species are of restricted range but the following are probably not: Circaetus fasciolatus (Nt), Zoothera guttata (Th), Sheppardia gunningi (Th), Anthreptes neglectus (Nt), and A. reichenowi (Nt). The estimations of breeding ranges are our own. Certain marginal species (of the not Threatened or Near-threatened group) were considered for inclusion as restricted-range species but were regarded just too widespread. Of these we would like to mention especially Olive-flanked Robin Cossypha anomala. Fülleborn's Black Boubou Laniarius fuelleborni (probably restricted-range if not conspecific with L. poensis) is not included as we follow Dowsett and Dowsett-Lemaire (1993) in regarding it conspecific with L. poensis. Malawi Batis Batis dimorpha (possibly restricted-range if given full species status as by a few authors) is not included since we follow e.g. Dowsett-Lemaire 1989) in not regarding it a full species.

2. Our list of forest bird species of special conservation importance in the Tanganyika-Nyasa Montane Forest Group may differ in certain respects from tables to be published for the Endemic Bird Area C24 in a coming BirdLife publication (The BirdLife Directory of Endemic Bird Areas, in prep.). The reasons are:

· Though the name of C24 may be changed from “Eastern Arc Mountains” (as it was incorrectly called in ICBP [1992]) to “Tanganyika-Nyasa Mountains” in the directory, the editors may choose to regard e.g. the lowland avifauna of the Usambaras as part of the Kenya-Tanzanian coastal forest EBA C23 (we have included them in C24) - it is difficult to distinguish between these biogeographical zones due to problems with overlap (the occurrence in the Usambaras of e.g. Tauraco fischeri, Otus irenae, Sheppardia gunningi, and Anthreptes reichenowi is probably best explained by affinity to coastal forests). Some authors believe that the S. Malawian mountains belongs to another biogeographical unit (e.g. Stuart et al. 1993).

· The coming directory from BirdLife may include only species of restricted range in their tables, using the text account for supplementary notes on Threatened and Near-threatened species not of restricted range, see (1).

· The editors of the directory may not follow the same taxonomy as we do (but may choose to follow e.g. Sibley and Monroe [1990] which will lead to the inclusion of some extra species).

Appendix 5.3. General distributions of the Threatened species, with notes on earlier Uluguru records.

General distribution of Scepomycter winifredae. An Eastern Arc forest endemic known only from four Tanzanian mountain ranges: the Ulugurus (type locality, discovered 1937, see below), the Ukagurus (discovered 1964 and recorded also on subsequent visits to the area, records are from between 1500 and 1850 m, described as common during a survey in 1978, heard moderately frequently during a survey in 1990), the Udzungwas (discovered 1982 but known only from Mwanihana Forest despite many other areas of the Udzungwas being worked; at Mwanihana, where it has been recorded from 1300 to 1700 m, it is fairly common) and the Rubehos (discovered 1993 when heard on several occasions by J. Fjeldså) (Collar and Stuart 1985, Evans and Anderson 1992 and 1993b, Jensen and Brøgger-Jensen 1992, Moyer 1993, Dinesen et al. 1993, Collar et al. 1994). All populations are assigned to the nominate race.

There were several earlier records from the Ulugurus, from the northern section as well as the Bunduki and Tchenzema areas (Moreau 1938 and 1946, Williams 1951, Friedmann and Stager 1964, Scharff et al. 1982, Stuart and Jensen 1985, and specimens in e.g. Copenhagen, Nairobi and Tring [Collar and Stuart 1985]). Stuart and Jensen (1985) described the species as being fairly common in the montane forest from 1350 to at least 2350 m. However, according to Moreau (1946) it is uncommon below 1650 m.

General distribution of Apalis chariessa. This very beautiful canopy species is - apart from the Ulugurus - only known from Southern Malawi (11 severely threatened montane forest patches, recorded between 500 and 1550 m, total population assessed to be not much above 100 pairs in the late 1980ies), Mount Chiperone in Mozambique (collected in 1950 at c. 1500 m, status unknown), the Udzungwa Mountains (discovered in the early 1980ies, recorded in five small and large forests, generally uncommon but locally more common, recorded from 1000 to 1600 m, with a single observation from 2000 m) and - very isolated - the lower Tana River, Kenya (where not recorded since 1960 and possibly extinct though still suspected to occur in some of the few remaining forest patches) (Benson 1950, Britton 1980, Collar and Stuart 1985, Dowsett-Lemaire 1989, Jensen and Brøgger-Jensen 1992, Dinesen et al. 1993, Lovett and Moyer in press, Moyer 1993, Collar et al. 1994, LAH and JOS pers. obs. from the Udzungwas 1994). Interestingly, A. chariessa has never been recorded in the Usambara Mountains.

All earlier Uluguru records of A. chariessa are from the Uluguru North F.R. (Stuart and Jensen 1985). A pair was collected in 1938 at 1100 m (Moreau 1940). The next published records are from the 1980ies. Several individuals were seen in mid July 1981 in Kinole Forest from 1250 to 1400 m (Stuart and Jensen 1981). There is one record from the western side in the forest above Morningside, 1500 m (D.C. Moyer pers. comm. to Stuart and Jensen 1985).

General distribution of Anthreptes rubritorques. A middle-altitude Eastern Arc endemic known from the Usambaras (discovered 1905, fairly common between 750 and 1200 m with flocks of up to 60 seen, less common between 1200 and 1500 m and with a single recent record from the foothills), the Ngurus (discovered in the 1940ies, foothills up to 1600m, apparently very uncommon although one report describes the species as common at flowering trees in the foothills), the Ulugurus (discovered in the 1950ies, apparently very uncommon, see below) and the Udzungwas (discovered 1982, very rare, 850-1550 m) (Collar and Stuart 1985, Jensen and Brøgger-Jensen 1992, Dinesen et al. 1993, Moyer 1993, Cambridge Tanzania Rainforest Project 1994, Collar et al. 1994).

The Uluguru population of A. rubritorques is documented only by five specimens, all collected by Andersen at Bagiro (Collar and Stuart 1985, Stuart and Jensen 1985). Two of these specimens were said to have been collected at 900 m and one as high as 1800 m (Britton 1978 and 1981, Collar and Stuart 1985) but Andersen's altitude data are often unreliable (Collar and Stuart 1985, Stuart and Jensen 1985, our Appendix 5.1) and it seems unlikely that he collected the species above 1600 m (Stuart and Jensen 1985).

General distribution of Malaconotus alius. This species - one of the rarest in Africa - is endemic to the submontane and montane forest of the Uluguru Mountains where recorded from 1300 to 2100 m (Collar and Stuart 1985, Stuart and Jensen 1985, Collar et al. 1994). Malaconotus alius was discovered in 1926 when two specimens were collected at 1830 m at Bagilo on the eastern slopes of the Uluguru North F.R. by Arthur Loveridge's native collector Salimu bin Asmani (Friedmann 1927, Loveridge 1960). The next observation is from 1948 when one bird was seen in forest above Bunduki on the western slopes of the mountains by J.G. Williams (Collar and Stuart 1985). Between 1952 and 1962 at least 13 specimens were collected, one being labelled "Ulugurus" and the remainder coming from Bagilo, 1800 m (Collar and Stuart 1985). Many of the Bagilo specimens were collected by Andersen and his collectors (Appendix 5.1). As indicated in Appendix 5.1, almost all Andersen’s specimens (of all species) from the Bagiro area were labelled 1800 m, and we therefore don't know precisely at which altitudes his M. alius specimens were collected. In 1981 the species was seen and heard on both sides of Lupanga Mountain (west side: immature at 1600 m, adult nearby at 1650 m plus apparently some extra records of birds only heard), and on the west escarpment of the Lukwangule Plateau at 2100 m (Stuart and Jensen 1981, Scharff et al. 1982). Another bird was seen at 1300 m on the eastern side of Lupanga in 1982 (Stuart and Jensen 1985). Several expeditions to the Ulugurus have failed to detect the species (Collar and Stuart 1985): The Cheney expedition searched diligently for the species during their 10 days of collecting above Bunduki in 1964 but failed to find it. They concluded that "It must be a bird of low numerical status and very local in its distribution to have evaded the search made for it" (Friedmann and Stager 1964). Another expedition to the Ulugurus in 1972 also failed to observe it (D.A. Turner pers. comm. 1977 to Collar and Stuart 1985).

General distribution of Ploceus nicolli. This rare Eastern Arc montane forest endemic is known only from three mountain ranges: the Usambaras (discovered 1931), the Ulugurus (discovered 1952) and the Udzungwas (discovered 1981); at all three localities it is elusive and occurs at low densities (Collar and Stuart 1985, Collar et al. 1994). In the East Usambaras the species has been seen at 900 m near Amani in the 1930ies but has possibly disappeared from this locality now (Collar and Stuart 1985). Another small population in the East Usambaras was discovered recently at 1250 m on the previously un-surveyed Mount Nilo (Collar et al. 1994). From the West Usambaras there are several records form between 1370 and 2200 m (several observations from both below and above 1700 m) (Collar and Stuart 1985). In the Udzungwas the species has been recorded from 1100 to 1700 m on the eastern escarpment, at 1850 and 2150 m in the Ndundulu Mountains (uncommon and not seen below 1850 m despite of a high intensity of general field observations) and from 1400 to 1570 m in the Nyumbanitu Mountains (seen at two localities) (Stuart et al. 1987, Jensen and Brøgger-Jensen 1992, Dinesen et al. 1993, LAH and JOS pers. obs. from the Nyumbanitu Mountains 1993-94).

The Uluguru population of P. nicolli is known from two specimens and one sighting (Collar and Stuart 1985, Stuart and Jensen 1985). One of the specimens (an adult male) was collected by Thorkild Andersen at 1500-1800 m in 1952 (Franzmann 1983) and the other (a female) was taken by G. P. Heinrich at 1600 m "near Morogoro" in 1961 (Ripley and Heinrich 1966). The sighting was of a bird at 1350 m in Kinole Forest in 1981 (Stuart and Jensen 1981). It appears to be a very low density bird in the Ulugurus (Collar and Stuart 1985, Stuart and Jensen 1985).

Appendix 5.4. General distributions of the Near-threatened species, with notes on earlier Uluguru records.

General distribution of Circaetus fasciolatus. A resident raptor of coastal woodlands and forests from Southern Somalia south to Natal and Zululand in South Africa (Britton 1980, Brown et al. 1982) but occurs also inland (see below). Its status in the coastal areas has been described as "normally uncommon, at best frequent, but probably commoner than supposed, as hard to locate and see" by Brown et al. (1982; covering its entire breeding range) and as "a reasonably common resident in forest (including remnants) throughout the coastal strip" by Britton (1980; covering East Africa). In Tanzania the species occurs inland in e.g. the Usambaras, the Ngurus, the Ulugurus, the Udzungwas and Ruaha N.P. (Britton 1980, Stuart and Jensen 1981, Brown et al. 1982, Fuggles-Couchmann 1984a, Stuart and Jensen 1985, Jensen and Brøgger-Jensen 1992). Other inland records include Voi in Kenya (Britton 1980) and Eastern Zimbabwe (Brown et al. 1982). C. fasciolatus is widespread and in no immediate danger but its habitat is declining (Collar and Stuart 1985).

From the Ulugurus there was only a single previous record (Kimboza Forest, 300 m) (Stuart and Jensen 1981 and 1985).

General distribution of Anthreptes neglectus. Known from several records from the Kenya coast and from several coastal forests in Tanzania and northern Mozambique (Collar and Stuart 1985). Occurs inland in certain Eastern Arc mountains, including the Ulugurus: Fairly common (at least in the foothills) in the East Usambaras where occurring from the foothills up to 1200 m (Britton 1980, Stuart 1983 and 1989, Cambridge Tanzania Rainforest Project 1994). In the Ngurus it is apparently scarce, records are from 1300 to 1500 m (Britton 1980). In the Udzungwas it has been recorded within every 100 m interval from 300 to 1300 m in Mwanihana Forest (Stuart et al. 1987, Jensen and Brøgger-Jensen 1992), in unknown densities at 750-1050 m and infrequently at 1400-1500 m in the southern part of the Udzungwa Scarp Forest Reserve (Jensen and Brøgger-Jensen 1992, Moyer 1993), and infrequently between 1350 and 1400 m in the Ndundulu and Nyumbanitu Mountains (Dinesen et al. 1993).

In the Ulugurus it is common in the eastern foothill forests (Stuart and Jensen 1985). It has been recorded on the eastern slopes up to 1300 m (Moreau in Stuart and Jensen 1985) and possibly 1800 m (Andersen in Stuart and Jensen 1985). There is also a record from "Uluguru" (Friedmann and Loveridge 1937). There are no definite records from the drier western slopes (Stuart and Jensen 1985).

General distribution of Nectarinia loveridgei. Known only from the Uluguru Mountains where it is common (Stuart and Jensen 1985, Collar and Stuart 1985). Earlier records are from between 1300 and 2350 m.

Relationship of N. loveridgei with other sunbirds. Several authors have commented on the close relationship of N. loveridgei with other members of the Nectarinia regia superspecies, especially Moreau's Sunbird N. moreaui and Eastern Double-collared Sunbird N. mediocris (e.g. Williams 1950, Hall and Moreau 1970, Stuart and van der Willigen 1980, Sibley and Monroe 1990, Dowsett and Dowsett-Lemaire 1993, Evans and Anderson 1993a). During our survey a large number of biometric measurements of N. loveridgei were obtained. These will hopefully be compared with similar measurements of N. moreaui (e.g. from the Ukagurus and the Rubehos where mistnetting have been carried out by other fieldworkers) and N. mediocris (from earlier extensive fieldwork in the Udzungwas by LAH and JOS) in a separate paper.

Appendix 5.5. Notes from our survey on forest species other than Threatened, Near-threatened and other restricted-range species.

Great Sparrowhawk (Black Sparrowhawk or Black Goshawk) Accipiter melanoleucus (G): Our record from Bunduki (where it was seen outside forest) is the second published from the Ulugurus. The first was from Bagilo (Stuart and Jensen 1985). In the Usambaras this species is stated to occur widely but in very small numbers, usually near forested areas (Stuart and Jensen 1985) and the same is the case in the Udzungwas (Jensen and Brøgger-Jensen 1992, LAH and JOS). This could also be the case in the Ulugurus.

African Goshawk Accipiter tachiro (P, T, G): Clearly the most common forest hawk in the Ulugurus. Territory calls were heard from many of our camps, typically in the mornings (Table 5.10). Recorded several times in the forest strip at Kigurunyembe. Stuart and Jensen (1985) surprisingly traced only two records from the Ulugurus but assumed that it is widespread. This has now been confirmed. We also saw the species in Singiza village (440 m).

Mountain Buzzard (Forest Buzzard) Buteo oreophilus (P, G): Frequently seen or heard soaring over the forest in the Kimhandu, Lanzi and Tchenzema areas. Stuart and Jensen (1985) also report the species to be common.

Crowned Eagle Stephanoaetus coronatus (G): This big monkey- and hyrax-eating eagle (normally easy to detect due to its far-carrying territory calls) was recorded at all three main localities in the Uluguru North and South F.R.s (and also in Kimboza F.R. and at Kigurunyembe), but comparing with other Eastern Arc montane forests visited by JF, LAH and JOS, it is our impression that we recorded it remarkably infrequently during the survey. Stuart and Jensen (1985) reported it to be common in the Kimboza Forest and in the forest of the northern section of the Ulugurus up to 2140 m. There were no earlier published records from the Uluguru South F.R.

Crested Guineafowl Guttera pucherani (G): Recorded in Kimboza F.R on several occasions during the visit by JF and JK and was also seen from the car while passing Kimboza. These are the first confirmed records from the Ulugurus since the late 1930ies, apart from an unidentified guineafowl feather found in 1984 (Stuart and Jensen 1985). The species may suffer from hunting in the Kimboza-Ruvu area (from e.g. local farmers and ruby miners) but no information was collected on the subject.

Buff-spotted Flufftail Sarothrura elegans (not recorded on this survey): This species, of which the only records from the Ulugurus are of birds collected by Loveridge and Andersen, was not recorded on this survey. It is most likely overlooked in the Ulugurus (Stuart and Jensen 1985). According to Urban et al. (1986) S. elegans tolerates a wide variety of habitats but is generally associated with forest or thick bush, favouring forest edge, clearings, second growth and more open types of forest, but also occurring in dense forest. It occupies a broad altitudinal range, from lowland rain forest to bamboo forest at 2600 m (Kenya) and Juniper/Podocarpus forest at 3200 m (Ethiopia).

Olive Pigeon Columba arquatrix (P, T, G): This large and beautiful pigeon of the canopy was very common at the localities we visited in the Uluguru North and South F.R.s, as is well illustrated by the plot data (Table 5.9). At all stations it was frequently heard singing or seen flying over in small flocks of typically three to five individuals. The largest flock recorded contained 42 individuals and was seen primo October. Flocks of up to seven individuals were frequently seen gathering in fruiting trees. At Kimhandu, 1520 m, single individuals and small flocks of 3-7 individuals were seen flying from a higher altitude to a lower just before dusk. It is the only species of pigeon we recorded at Kimhandu-5 (2520 m) and on the Lukwangule Plateau.

C. arquatrix was surprisingly common in the Ulugurus compared to densities found in the Udzungwa Mountains 1991-92 by LAH and JOS (six months of fieldwork July-April between 1350 and 2400 m in the Ndundulu and Nyumbanitu Mts. and Kisinga-Rugaro F.R.). The species is known to be rather migratory in other parts of its range (see e.g. Dowsett-Lemaire [1989]), so possibly its abundance in Eastern Arc forests fluctuates from year to year, linked to availability of fruiting trees, but little is known about this. The difference between the Ulugurus and the Udzungwas may also be “permanent”.

Bronze-naped Pigeon Columba delegorguei (P, T, G): Heard singing frequently near many stations but it was absolutely not as abundant as C. arquatrix at the localities we visited. C. delegorguei appears to avoid high altitude forest where C. arquatrix was seen frequently (Table 5.4).

Lemon Dove Aplopelia larvata (M, P, T, G): This shy ground haunting pigeon was recorded very frequently only at Lanzi-3 (2110 m) where four individuals were mistnetted and where several individuals were seen escaping from our mistnets and several flushed from the ground. In the other study areas it appeared to be scarce. See also Appendix 5.8.

Tambourine Dove Turtur tympanistria (P, G): We heard the species now and then but nowhere very frequently except in the forest strip at Kigurunyembe where it was common. This contradicts the description given in Stuart and Jensen (1985), saying that the species is fairly common throughout the Uluguru forests, from the foothills up to at least 2100 m. Most of our records at the main localities are from near the Tegetero-1 (1345 m) camp but despite that it has a higher plot data score than Columba delegorguei at this station (Table 5.9), it is our clear impression from the general field observations that C. delegorguei was the most common of the two also here. Turtur tympanistria is probably most common at low and intermediate altitudes.

Livingstone’s Turaco Tauraco livingstonii (P, T, G): Common throughout the Uluguru North and South F.R.s as is well illustrated by our plot and tape recording data (Tables 5.9 and 5.10). The species, who’s far-carrying, harsh voice is one of the typical morning sounds in the forest, is represented on the dawn tape recordings at all stations. Heard in some small forest patches we passed in the open country at 1500-1700 m and is quite common even in small patches of exotic trees in the Mgeta, Nyiandira and Tchenzema areas.

Barred Long-tailed Cuckoo Cercococcyx montanus (P, T, G): The characteristic and far-carrying song of this species was heard only occasionally at Kimhandu and Lanzi. At Tegetero area we heard it quite frequently, even during the nights. It is our impression, however, that it was less common also at Tegetero than what was found during fieldwork in the Udzungwas 1991-92 (unpublished records from the Ndundulus which are forested from 1340 to 2400 m, the species was most common up to c. 1750 m in November-January, LAH and JOS). However, Stuart and Jensen (1985) describe it as abundant in the Ulugurus above 1500 m in the hot season. See also Appendix 5.8.

Yellowbill (Green Coucal) Ceuthmochares aereus (not recorded on this survey): We did not find this species which has been recorded from "East Uluguru" at 400 m, Bunduki, Mkaraji and "Uluguru" up to 1800 m. Stuart and Jensen (1985) reported this species to be uncommon in the Ulugurus, compared to the Usambaras. According to Fry et al. (1988) the species usually occurs in secondary forest and gallery forest, occasionally riparian Acacia.

Barred Owlet Glaucidium capense (not recorded on this survey): The only earlier record of this small owl is from Kibungo Forest (close to or synonymous with Kimboza F.R.). Britton (1980) describes the species as a local and generally uncommon resident in forest and richer woodland, mainly below 1200 m.

Wood Owl Strix woodfordii (T, G): Heard at night from most of our camps.

Unidentified owl (G): Two independent observations, by two persons, of an unidentified (buffy brown?) big owl were made on 5 October. The bird was seen gliding through the forest near the Kimhandu-2 (1710 m) station. Further visitors should be alert of the possible presence of a Bubo sp. in the area. No other owls than Strix woodfordii were heard at night time, however. A Finnish ornithologist who has been working around the Bagilo area, reports that he has found a feather from an owl which was not S. woodfordii (letter to Neil Burgess). These two findings stress the need for further searches for owls in the Uluguru forests.

Scarce Swift Schoutedenapus myoptilus (P, T, G): Schoutedenapus myioptilus, which had never been recorded in the Uluguru Mountains before, was heard and seen frequently over forest at Kimhandu, Lanzi, Tegetero and Tchenzema. It was scored on plots at Tegetero-2 (1535 m) and Tchenzema 2150 m. The birds foraged over the forest alone, pair wise or in flocks of up to at least five. On several occasions they mixed up with flocks of up to 150 swallows, especially Red-rumped Swallow Hirundo daurica, flying in front of a depression. Many were also seen at Kigurunyembe at 600-1000 m near the forest strip, mixing up with Little Swift Apus affinis and Black Roughwing Swallow Psalidoprocne pristoptera. Within the Eastern Arc this species has otherwise only been recorded in the Usambaras, where it was discovered in 1980 (Stuart and Turner 1980) and the Udzungwas where it was discovered 1982 (Stuart et al. 1987; heard and seen relatively frequently during fieldwork in forest in the Udzungwas by LAH and JOS 1991-92).

In the Malawian forest-capped mountains S. myoptilus is a breeding migrant (settling in October or early November and common until March when they suddenly disappear) which presumably breeds in trees in the montane forests (at least in the north) (Dowsett-Lemaire 1989). We have preferred to regard it a forest species although it has not yet been documented whether it is restricted to forested areas in the Eastern Arc.

Narina Trogon Apaloderma narina (M, G): During this survey recorded only in the Kimboza F.R. where several individuals were seen and a single mistnetted. Stuart and Jensen (1985) reported the species to be common in the eastern foothills. There are a few earlier records of A. narina from the Uluguru North and South F.R.s from altitudes at which A. vittatum is common (Stuart and Jensen 1985) but in general these two species are altitudinally segregated in the Ulugurus as is the case also in other forests holding both species.

Bar-tailed Trogon Apaloderma vittatum (M, P, T, G): Reported to be found fairly common throughout the montane forests (Stuart and Jensen 1985) and was seen and heard frequently during this survey. Recorded on plots at all stations except Kimhandu-3 (1940 m) and Kimhandu-5 (2520 m). Also seen in Singiza village (440 m) in mango trees near houses.

Crowned Hornbill Tockus alboterminatus (G): No previous surveys in the Ulugurus have recorded T. alboterminatus in the submontane and montane forests of the Ulugurus (except an old record from 900 m, probably at the forest edge), and we also failed to do so. The species seems to occur only in Kimboza F.R., where it was seen frequently during this survey (also in village/woodland edge), and in Mvuha F.R. and other dense woodlands in the foothills where we saw it from the car while passing. There is, though, an earlier record from Mgeta (Baker pers. comm. to Stuart and Jensen 1985).

The altitudinal abundance pattern of T. alboterminatus recorded in the Ulugurus differs from what LAH and JOS found in the Ndundulu Mountains (forest cover from 1350 to 2400 m, in the Udzungwas) during four and a half months of fieldwork 1991-92: At this locality T. alboterminatus was recorded frequently deep inside submontane and montane forest, in the dry season as well as in the hot season and records were made up to 2200 m. It was, however, clearly less common than Ceratogymna brevis. At Mwanihana on the eastern scarp of the Udzungwas it has been recorded only up to 700 m (Stuart et al. 1987). Probably it cannot be excluded that the abundance of T. alboterminatus fluctuates from year to year, following abundance of fruits. Fry et al. (1988) characterise T. alboterminatus as an arboreal species found in montane, coastal, riparian and secondary forest patches, also extending to some areas of dense woodland; occurs up to 3000 m.

Trumpeter Hornbill Ceratogymna bucinator (G): Seen frequently in Kimboza Forest during this survey. Also Stuart and Jensen (1985) report it to be common only in Kimboza. The only other record from the Ulugurus is from Bagilo (Friedmann and Loveridge 1937, Stuart and Jensen 1985). According to Fry et al. (1988) it inhabitants interior and edge of montane, riparian and coastal forest, also moist woodlands and mangroves. It may be overlooked in the lower parts of the submontane areas of the Ulugurus.

Silvery-cheeked Hornbill Ceratogymna brevis (P, T, G): This big and noisy canopy bird was common at all localities and was recorded on plots at seven of 13 stations (up to 36 % of the plots per station). It was not, however, recorded above 2160 m. Seen in small groups of trees outside forest (although not very far from this) in the Bunduki area. There were no earlier published records from the southern section.

White-eared Barbet Stactolaema leucotis (G): Recorded very frequently in Kimboza F.R. during the visit by JF and JK. Also Stuart and Jensen (1985) reported it to be common there. According to Britton (1980) the subspecies leucogrammicum occupies highland areas to the south, from the Ulugurus to Mahenge, but as is evident from Table 5.4 there are actually no records from submontane and montane areas of the Ulugurus (Mkaraji is at c. 500 m altitude). According to Fry et al. (1988) S. leucotis frequents large trees in forest and along streams, usually in hills and mountains (to 2600 m, Arusha area) in the north, to coast south from S Mozambique. A slow and monotone typical tinkerbird/barbet call heard at Lanzi, 1960-2000 m, on 8 October could have been from this species.

Green Barbet Stactolaema olivacea (P, G): During this survey we did not record it frequently, except in the lowest part of the Tegetero forest where it was noted frequently in mixed feeding parties during the general field observations, and in the forest strip at Kigurunyembe. Stuart and Jensen (1985) reported it to be most common on the eastern slopes in the northern section of the Ulugurus.

Eastern Green Tinkerbird Pogoniulus simplex (not recorded on this survey): Although Stuart and Jensen (1981 and 1985) report it to be common in the Kimboza Forest it was not recorded there during our survey. Occurs up to 900 m in the Usambara Mountains (Stuart and Jensen 1985, Evans and Anderson 1992b), and up to 1500 m in Malawi (Fry et al. 1988), leaving a theoretical possibility that it could be found in small numbers in lower parts of the Uluguru North.

Moustached Green Tinkerbird Pogoniulus leucomystax (M, P, T, G): This species has a penetrating monotypic call which is audible on long distance, and was heard frequently at all localities as illustrated by the plot data (Table 5.9). At Tchenzema it was recorded on no less than 92 % of 25 plots. Everywhere it was, however, hard to get visual observations of this small barbet. Judging from how frequent vocalizations were heard during general field observations and further supported by the plot data, P. leucomystax is more common in the Kimhandu area and at Tchenzema than at Lanzi and Tegetero. It could be that breeding activities had reduced the vocal activity at Lanzi and Tegetero, but the difference may also have to do with a difference in the availability of certain fruits.

The Uluguru population of P. leucomystax has two different calls/songs. The deepest of these (heard frequently) has never been heard during several months of fieldwork carried out in the Udzungwas by LAH and JOS.

Golden-rumped Tinkerbird Pogoniulus bilineatus (P, G): Not recorded at Lanzi and much less common than P. leucomystax at Kimhandu and Tegetero. Also Stuart and Jensen (1985) remarked that the species is clearly outnumbered by P. leucomystax at intermediate elevations. At Kigurunyembe it was common in the forest strip. In Kimboza Forest it is reported to be “not uncommon” but outnumbered by P. simplex (Stuart and Jensen 1985).

Scaly-throated Honeyguide Indicator variegatus (G): Heard on a single occasion during our survey (Tegetero area). Stuart and Jensen (1985) suggest that the species is overlooked in the Ulugurus, and that it probably occurs widely throughout the forest in small numbers.

Eastern Least Honeyguide (Pallid Honeyguide) Indicator meliphilus (not recorded on this survey): Stuart and Jensen (1985) mention a specimen "almost certainly of this species rather than the Kilimanjaro Honeyguide Indicator narokensis" collected by Andersen in forest at 900 m. In the Usambara mountains it occurs in small numbers in lowland forest, therefore Stuart and Jensen (1985) suggest that it can also be expected to occur in Kimboza Forest and elsewhere in the foothills.

Olive Woodpecker Dendropicos griseocephalus (P, G): Found near most of our stations but at low densities. Recorded in many of the mixed feeding parties we saw in the Tegetero area.

African Broadbill Smithornis capensis (M, P, T, G): Recorded most frequently at Tegetero with a few observations from near Kimhandu-2 (1710 m). Not recorded at Lanzi. Compared to densities found in the Udzungwas in 1991-92 (LAH and JOS, unpublished records from fieldwork in the Ndundulus above 1350 m, where the species was common up to 1750 m), S. capensis was surprisingly uncommon at the localities we visited. The low density is underlined by the few individuals audible on our tape recordings (the voice of this species is one of the most typical "morning sounds" in areas where it is abundant). Also Stuart and Jensen (1985) considered the species to be less common in the Ulugurus than in the Udzungwas (and the Usambaras). The rarity may be a result of the small area in the Ulugurus of premontane forest which is the source habitat for this species. See also Appendix 5.8.

Purple-throated Cuckoo-shrike Campephaga quiscalina (G): A female was seen in Kimboza F.R. It is the rare subspecies C. q. muenzneri that occurs in the Ulugurus (Britton 1980).

Grey Cuckoo-shrike Coracina caesia (P, G): In the Kimhandu area the only records were of a few individuals seen and heard on some occasions near the Kimhandu-3 (1940 m) station. It is a silent bird which can be overlooked if not in a feeding party (the most commonly heard voice is a high-pitched, low-volumed short whistle - not what one would expect from a bird of its size) but it must be regarded as uncommon in the Kimhandu area. In the Lanzi area it was seen on some occasions in a wider altitudinal range but also here it appeared to be uncommon. In the Tegetero area it was common below 1600 m, and it was seen in most of the feeding parties studied. See also Appendix 5.7.

Shelley's Greenbul Andropadus masukuensis (M, P, T, G): Recorded very frequently at most stations, being absent only at the high-altitude station Kimhandu-5 (2520 m) as illustrated by the plot and mistnetting data. The second most frequently mistnetted species.

Uluguru Mountain Greenbul Andropadus neumanni (Mountain Greenbul Andropadus tephrolaemus by some treated as an endemic subspecies only A.t. neumanni) (M, P, T, G): Very common in the Uluguru North and South F.R.s (see mistnetting, plot and tape recording data). The plot data show that at Kimhandu and Lanzi the species is most common above 1700 m altitude. At Lanzi it is common even down to 1700 m. At Kimhandu-5 (2520 m) it is very common in the elfin forest and the only greenbul recorded. Also very common in elfin forest on the Lukwangule Plateau. Morphologically this endemic (sub)species is one of the most distinct in the A. tephrolaemus complex. The repertoire of calls and song (tape recorded during our survey) differ slightly from the vocalisations we have heard from the subspecies occurring in the Udzungwas (LAH and JOS).

“Stripe-cheeked” Greenbul Andropadus olivaceiceps (M, P, T, G): Relatively common, being absent only at the highest altitudes. Recorded most frequently in the Tegetero area below 1500 m. See also Appendix 5.8.

Little Greenbul Andropadus virens (M, P, G): Very common in Kimboza FR. Common at Kimhandu-1 and -6 (1520 and 1540 m) (see mistnetting and plot data) which were situated less than 500 m from the lower forest edge. Recorded in Eucalyptus forest at Bunduki, in the forest strip at Kigurunyembe (common) and at 1500 m at Morningside. The above-mentioned combined with the fact that the species was not recorded higher than at 1200 m in the Tegetero area, shows that in the Uluguru North and South F.R.s it is the presence/absence of edge habitat rather than of submontane forest that determines its abundance.

Yellow-bellied Greenbul Chlorocichla flaviventris (not recorded on this survey): This species, of which the only Uluguru records are of a specimen collected by Loveridge (1933; exact locality unknown) and of a bird collected in Kibungo Forest by Moreau, is either unexpectedly uncommon or greatly overlooked in the Ulugurus (Stuart and Jensen 1985). According to Britton (1980) this is a common species of eastern lowlands, frequenting forest, forest edge, woodland and bush land thicket and secondary scrub habitats, regularly as high as 1700-2100 m in some northern parts of its range. Keith et al. (1992) furthermore mention occurrences up to 1700 m in Malawi but also describes it as a species mainly of lowland forest.

Terrestrial Greenbul Phyllastrephus terrestris (M): A single bird was mistnetted in Kimboza Forest during our survey. The only earlier observation from the Ulugurus is of a specimen collected by Loveridge (Loveridge 1933, Stuart and Jensen 1985). Elsewhere in eastern Tanzania this is an uncommon species of lowland forest and thicket (Britton 1980, Stuart and Jensen 1985).

Fischer's Greenbul Phyllastrephus fischeri (G): Several were heard in the forest strip at Kigurunyembe (600-850 m) during the visit by JF and JK. Stuart and Jensen (1985) report this species to be common in both Kibungo and Kimboza Forests. Our records are the first from outside the foothills of this species, which is mainly restricted to lowland forest. The highest altitudinal record listed for the species in Britton (1980) and Keith et al. (1992) is 600 m (in the Usambaras).

Grey-olive Greenbul Phyllastrephus cerviniventris (M, G): Two birds were mistnetted at Kigurunyembe where the species was also heard. There were only two earlier records from the Ulugurus (Stuart and Jensen 1985): one of a specimen from “Uluguru” (Friedmann and Loveridge 1937) and one of two birds collected at Morningside by K.M. Howell in 1982 (Stuart and Jensen 1985). According to Stuart and Jensen (1985), this species has only been found in thickets away from the forest in the Ulugurus and the Usambaras. Britton (1980) describes it as local and uncommon in forests and streamside thickets up to 1500 m, with a curiously fragmented distribution.

Placidus Greenbul Phyllastrephus placidus (Olive Mountain Greenbul) (M, P, T, G): Common in the submontane and montane areas except at the highest altitudes. Monospecific feeding parties of this species were seen in the Uluguru South and North F.R.s.

Yellow-streaked Greenbul Phyllastrephus flavostriatus (P, G): This typical mixed feeding party species was recorded in almost all drongo parties studied at Tegetero and in Kimboza Forest (see Appendix 5.7). The highest altitude in which we noted drongo parties was 1450 m. A few P. flavostriatus were seen between 1450 and 1500 m, moving solitarily (few singing birds were heard) or in small parties with e.g. Dark-backed Weaver Ploceus bicolor.

There are no previous records of P. flavostriatus from the Uluguru South F.R. and we also failed to record it there during this survey. The species probably depends on source habitat in the submontane zone (below 1500 m) although we have recorded it up to 2000 m in the Udzungwas (LAH and JOS pers. obs., from the Ndundulus 1991-92 where the species was uncommon above 1750 m). Stuart and Jensen (1981 and 1985) report it to be common from 300 m in the eastern foothills (probably the Kimboza Forest) up to 1500 m. The earlier record from 1800 m (Table 5.4) is one of Andersen’s and may be from a lower altitude.

Tiny Greenbul Phyllastrephus debilis (G): Only one individual was seen during this survey, namely in Kimboza Forest. Stuart and Jensen (1985) describe this species as being common in the eastern foothill forests.

Red-tailed Ant Thrush Neocossyphus rufus (M, G): Two individuals were mistnetted in Kimboza Forest where also visual records were made. Stuart and Jensen (1985) reported the species to be common in the eastern foothill forests. There are earlier records of the species from Morningside, 1500 m (Baker pers. comm. to Stuart and Jensen 1985), and from “Uluguru” (Friedmann and Loveridge 1937). The species is said to follow swarms of driver ants Dorylus sp. but this is not confirmed by observations in Kimboza (this survey) and other Tanzanian forests as it is usually seen feeding on insects 5 m above ground or higher, mostly sally-gleaning and perch-gleaning, and often flycatching (JF).

Olive Thrush Turdus olivaceus (M, P, T, G): It is our clear impression that this species is more common in the Ulugurus than in the Udzungwas (LAH and JOS). The mistnetting catch rates are quite similar to those of the Orange Thrush Zoothera gurneyi but T. olivaceus was seen and heard much more frequently than Z. gurneyi (see plot data).

Orange Thrush Zoothera gurneyi (M, P, T, G): According to Stuart and Jensen (1985) the species avoids the high altitude forest in the Usambaras and they speculate whether the same might be the case in the Ulugurus. The results of the present survey show that it is common throughout the forests, even at high altitudes. There were no earlier published records from the Uluguru South F.R. (Stuart and Jensen 1985). See also Appendix 5.8.

Starred Robin (White-starred Robin) Pogonocichla stellata (M, P, T, G): Common at Kimhandu, Lanzi, Tegetero and Tchenzema during our survey, occurring along the entire gradient, as illustrated by plot, mistnetting (third-most frequently mistnetted species) and tape recording data. From the general field observations it was our impression that the species was seen more rarely around Tegetero-1 (1345 m) (the lowest of our stations) but this is not supported by the mistnetting or the plot data. We recorded the species also in tiny patches of wattle (Acacia) and bushes below the forest, e.g. in the Tchenzema area at c. 1700 m where several individuals were seen. See also Appendix 5.8.

Olive-flanked Robin Cossypha anomala (M, P, T, G): Common above 1900 m (at Lanzi also down to 1700 m) and a characteristic bird of higher altitudes (see mistnetting and plot data). Tame and on several occasions seen close to tents or near mistnets following human activities (putting up mistnets or establishing a new camp) that expose the bare ground under the leaf litter. Heard in some very small forest patches we passed in the open country (less than 500 m from “the main forest”) at 1500-1700 m altitude.

Red-capped Robin Cossypha natalensis (M, G): Common in Kimboza F.R. where mistnetted in good numbers during the visit by JF and JK (seven of the nineteen birds ringed). Also recorded frequently in the forest strip at Kigurunyembe where four individuals were mistnetted (Table 5.12). Stuart and Jensen (1985) describe the species as very common in the eastern foothills. There are also earlier records from "Uluguru" and Bagilo (Friedmann and Loveridge 1937) and from Bagilo and "Uluguru", 1500-1800 m (Andersen in Stuart and Jensen), showing that the species occurs also at intermediate elevations. It was not recorded at Tegetero during our survey, supporting Stuart and Jensen’s (1985) assumption that it is much less common at intermediate elevations that in the lowland forests.

Evergreen Forest Warbler Bradypterus mariae (M, P, T, G): Common at all altitudes in the Uluguru North and South F.R.s (e.g. Table 5.9). At high altitudes, e.g. at Kimhandu-5 (2520 m) and at the Lukwangule Plateau, it occurs sympatrically with Cinnamon Bracken Warbler B. cinnamomeus (for the latter, see Appendix 5.6). Heard in several very small forest patches passed in the open country (less than 500 m from "the main forest") at 1500-1700 m altitude. Some of the birds recorded at Bunduki were observed in Eucalyptus forest (in bracken patches). At Tchenzema the species was seen at several occasions in small thickets around the shambas and a single bird was even seen in weeds in a maize field. See also Appendix 5.8.

Kretschmer's Longbill Macrosphenus kretschmeri (G): The only records made during this survey are from Kimboza Forest where several individuals were heard. Apart from the eastern foothill forests earlier records are from Kinole Forest, 1100 m (Moreau in Stuart and Jensen 1985), the northern slopes of Lupanga at 1000 m (Stuart in Stuart and Jensen 1985) and Bagilo, 1800 m (Andersen in Stuart and Jensen 1985). According to Stuart and Jensen (1985) M. kretschmeri appears to be considerably less common in the Ulugurus than in the Usambaras and the Udzungwas. Britton (1980) describes the species as wide-ranging but generally uncommon in forest and fringing thickets, virtually endemic to E Tanzania.

Brown Woodland Warbler Phylloscopus umbrovirens (endemic subspecies fugglescouchmani) (M, P, T, G): Phylloscopus umbrovirens has its southernmost population in the Ulugurus, the nearest population to the north being in northern Tanzania more than 350 km away. We recorded it frequently around our Kimhandu-5 (2520 m) camp in the elfin forest where it is clearly one of the character species as illustrated by the mistnetting and plot data. It is also common in the Lanzi area above 2200 m and on the Lukwangule Plateau above Tchenzema. It is altitudinally segregated above its close relative P. ruficapillus with a contact zone at c. 2200 m. We did not record P. umbrovirens in the Uluguru North F.R., doubtless only because we did not survey the uppermost parts there. The earlier record from 900 m is of a specimen collected by Andersen (Stuart and Jensen 1985).

Yellow-throated Warbler Phylloscopus ruficapillus (M, P, T, G): Recorded frequently at all three main localities but never above 2160 m. See above under P. umbrovirens.

Bar-throated Apalis Apalis thoracica (endemic subspecies uluguru) (M, P, T, G): Very common at all three main localities (see mistnetting, plot and tape recording data). Generally uncommon in the lower part of the forest (e.g. the area around the Tegetero-1 1345 m) but becomes one of the most characteristic species at higher altitudes. Often watching curiously at 1-2 m distance and must be considered to be more tame in the Ulugurus than in e.g. the Udzungwas (where JF, LAH, JK and JOS have experience with it).

Black-headed Apalis Apalis melanocephala (P, G): This species was recorded only in the Uluguru North F.R. where it was seen in almost all mixed feeding parties encountered. There are no earlier records from the Uluguru South F.R. and it must be considered to be absent or occurring at only very low densities in that section which does not offer the right conditions - submontane and montane populations of A. melanocephala are normally confined to altitudes below 1700 m, primarily the submontane belt where it is most frequently recorded in mixed feeding drongo parties (such parties are lacking in the Uluguru South F.R.).

Green-backed Camaroptera (Bleating Bush Warbler) Camaroptera brachyura (P, T, G): Recorded on several occasions in Kimboza F.R. during the visit by JF and JK. At Kigurunyembe it was common in the forest strip and also in the park around Teachers College, one was mistnetted. Heard often in the disturbed shrubby forest near the lower edge at Kimhandu which is reflected in the plot data from Kimhandu-1 and -6 (1520 and 1540 m). At Tegetero we recorded it only by the standard tape recordings at Tegetero-1 (1345 m).

Dusky Flycatcher Muscicapa adusta (M, P, G): Generally uncommon but was seen occasionally at most localities. Usually seen sitting on a branch in the outermost trees along forest glades or in trees along rivers.

Ashy Flycatcher Muscicapa caerulescens (G): The single record from this survey was made in surrounding cultivation in the Ukwama area. Stuart and Jensen (1985) report the species to be common in the eastern foothill forests. At higher elevations in the Ulugurus it is a species of plantations and other wooded areas rather than forest (Stuart and Jensen 1985). Britton (1980) describes the species as wide-ranging up to 1800 m, in forest edges and glades, riparian forest strips and richer woodland, where it is seldom common.

Forest Batis (Short-tailed Batis) Batis mixta (M, P, T, G): Common at most stations (see plot and mistnetting data). It was, however, not recorded at the highest altitudes.

Black-throated Wattle-eye Platysteira peltata (G): A few birds were heard in the forest strip at Kigurunyembe. Earlier records are from Morningside, 1200 m (Stuart and Jensen 1985), “Uluguru” (Friedmann and Loveridge 1937) and Bagilo (Andersen in Stuart and Jensen 1985). Britton (1980) describes the species as being widespread but generally uncommon in forest patches, forest edges, bush land thickets and gardens up to 3000 m.

Little Yellow Flycatcher Erythrocercus holochlorus (G): A few individuals were seen in Kimboza Forest during our survey. Stuart and Jensen (1985) report it to be common in the eastern foothill forests.

White-tailed Crested Flycatcher Trochocercus albonotatus (M, P, T, G): This small fan-tail flycatcher, usually very vocal and active, is very common in the Uluguru montane forests (see, e.g. plot data). It regularly forages in the understorey down to a level of about two-three feet and is therefore mistnetted at most of the stations. However, the best estimate of its abundance clearly comes from the plot data. See also Appendix 5.8.

Blue-mantled Flycatcher (Crested Flycatcher) Trochocercus cyanomelas (M, G): Two birds were mistnetted in Kimboza Forest where it was also seen frequently. Fairly common in the forest strip at Kigurunyembe where one was mistnetted (Table 5.12). Stuart and Jensen (1985) report it to be common in the eastern foothill forests at 300 m. Apart from the foothills there are earlier records also from Bunduki, 1500-1700 m, Mkaraji, "Uluguru", 900 m, and "Uluguru" (Stuart and Jensen 1985).

Paradise Flycatcher Terpsiphone viridis (P, T, G): As illustrated by the plot data the density was highest at the two lowest Tegetero stations (1345 and 1535 m). In this area it was a frequent member of mixed feeding parties. Relatively common at the lowest altitudes in the Kimhandu and Lanzi areas. One individual mistnetted in the forest strip at Kigurunyembe at 700 m. T. viridis is not restricted to forest habitats and was seen in mango trees near small villages on 30 October on our walk from Ukwama village to Kolero village. See also Appendix 5.8.

Pale-breasted Illadopsis Illadopsis rufipennis (M, G): Two birds were mistnetted in Kimboza Forest where the species was also heard. Stuart and Jensen (1985) describe the species as being common in Kimboza. It is interesting that this species has not been recorded above 900 m in the Ulugurus. In the Udzungwas the species can be found up to 1900 m although it is uncommon above 1700 m (Jensen and Brøgger-Jensen 1992, LAH and JOS) and in the Usambaras it occurs up to 1200 m.

African Hill Babbler Alcippe abyssinica (M, P, T, G): Quite common locally, e.g. at the Lanzi-3 (2110 m) station and at Tchenzema. At Tegetero and partly Kimhandu it was not recorded frequently and it is surprising that we did not record it above 2000 m at Kimhandu. There may be local variation in its abundance but it is our impression that it is less common in the Ulugurus than what was found during five months of fieldwork in the Udzungwas 1991-92 (LAH and JOS). According to Stuart and Jensen (1985) it is generally fairly common and tends to be most numerous at the higher elevations of the southern section occurring up to at least 2500 m. The species was seen in small forest patches near Tchenzema at c. 1700 m.

Olive Sunbird Nectarinia olivacea (M, P, T, G): At the localities visited during this survey it was recorded frequently only in the lowest part of the Tegetero area (21 mistnetted, Table 5.6), in Kimboza Forest (one mistnetted, Table 5.11) and in the forest strip at Kigurunyembe (one mistnetted, Table 5.12). At Tegetero-1 (1345 m) this species was actually the most frequently mistnetted bird (88 % of all individuals ringed) and exceeded N. loveridgei in numbers. In the Uluguru South F.R. it was only common in some areas near the lower forest edge (three mistnetted at Kimhandu-1, Table 5.6). Seen also far from forest in well-wooded cultivated land when we moved to and from the forests (e.g. several were heard while moving by car from the lowlands to the Tegetero area). Seen in bushy land in the Bunduki area at c. 1300 m.

Yellow White-eye Zosterops senegalensis stierlingi and Z. s. andersoni (M, P, G): Relatively common at all three main localities as illustrated by the plot data. Stuart and Jensen (1985) report it to be common in the montane forest and much less common in the lowland forests. Birds with characters as stierlingi were seen in small forest patches near Tchenzema mission at c. 1700 m.

The birds seen in the mountains during this survey showed the characteristics of the form Z. s. stierlingi. Moreau (1957) noted the same. He furthermore noted that the birds seen in the lowland forest showed the characteristics of the lowland form Z. s. anderssoni. White-eyes were not seen in the Kimboza during our survey, unfortunately. Further research is necessary to clarify the taxonomy of the white-eye complex. Z. s. stierlingi may prove to be a paraspecies in the group of montane white-eyes (Z. poliogaster sensu latu) found only in the montane forests, and replaced by Z. senegalensis (anderssoni) in the intervening lowlands.

Green-headed Oriole Oriolus chlorocephalus (P, G): This very beautiful oriole was recorded on several occasions in the lower parts of the Tegetero area in the Uluguru North F.R. and in Kimboza Forest. Our record from Tchenzema on the Lukwangule Plateau, where a single bird was seen and heard, is the first published from the Uluguru South. Britton (1980) gives the maximum altitude for this species as 1800 m (probably based on an Andersen record from the Ulugurus) and our record from 2500 m is therefore from an unusually high altitude.

Fülleborn's Black Boubou Laniarius fuelleborni (M, P, T, G): Relatively common throughout the Uluguru North and South F.R.s and one of the characteristic birds of higher altitudes (see plot data). The highest density was recorded at Tchenzema. Heard in some small forest patches and scrubs passed in open country (up to 500 m from "the main forest") at 1500-1700 m and in forest patches on the Lukwangule Plateau.

Black-fronted Bush Shrike Telophorus nigrifrons (Many Coloured Bush Shrike Telophorus nigrifrons [nigrifrons]) (P, T, G): Very common in the Tegetero area where it was seen in almost all mixed species parties encountered and outside mixed species parties too. The species was surprisingly uncommon in the Kimhandu and Lanzi areas (apparently more common at Lanzi than at Kimhandu). Stuart and Jensen (1985) mention records from Kimboza and Kibungo Forests in June and suggest that a part of the population moves down to this locality in the cold season. See also Appendix 5.8.

Four-coloured Bush Shrike Telophorus quadricolor (not recorded on this survey): Earlier Uluguru records are from "the Uluguru Mountains" and “the Ulugurus” (Loveridge 1933, Britton 1980, Swynnerton in Stuart and Jensen 1985). The taxon is thinly distributed in forest undergrowth and thicket throughout the coastal lowlands, occurring also inland at some localities (Britton 1980).

Nicator Nicator gularis (G): Recorded by JF and JK in Kimboza F.R., at Bunduki and at Kigurunyembe (a few territories). The only earlier records from the Ulugurus are from Kibungo Forest (Moreau in Stuart and Jensen 1985), the northern slopes of Lupanga at 1000 m (Stuart and Jensen 1985), Nyingwa (Friedmann 1928) and “Uluguru” (Swynnerton in Stuart and Jensen 1985). The species apparently occurs at only low densities in the Ulugurus. Concerning its general distribution Britton (1980) mention that it is not restricted to forest but also occurs in bush land and woodland thicket.

Square-tailed Drongo Dicrurus ludwigii (P, T, G): Common in the Uluguru North F.R. below 1450 m and in the Kimboza F.R. At these two localities it was invariably seen in the mixed feeding drongo parties which were a prominent phenomenon at these two localities (see Appendix 5.7). In the forest strip at Kigurunyembe it was fairly common but on territories and not seen in connection with mixed feeding parties. In the Uluguru South F.R. it was recorded only on a single occasion (near the Kimhandu-6 [1540 m] station; a pair was seen and heard on a single occasion; there were no signs of feeding party activities when the birds were seen). The almost complete absence of D. ludwigii from the southern section should probably be explained by the general absence of submontane habitats in this area.

Waller's Red-winged Starling Onychognathus walleri (P, T, G): This canopy species was common at all three main localities. Plot data supports the impression from the general field observations, that it was most common around the Lanzi-3 (2110 m) station.

Black-breasted Glossy Starling Lamprotornis corruscus (not recorded on this survey): The only records of this species from the Ulugurus are of a few birds seen in Kimboza F.R. in November 1984 (Stuart and Jensen 1985). Generally a common and adaptable bird of forested and formerly forested areas up to at least 1400 m (Britton 1980).

Dark-backed Weaver Ploceus bicolor (P, G): Clearly one of the core species of the mixed feeding drongo parties in the submontane areas at Tegetero and in the lowland Kimboza Forest (see Appendix 5.7). Several pairs were seen in the forest strip at Kigurunyembe. We suppose that the species is (at least almost completely) absent from the Uluguru South F.R. because of the lack of submontane source habitat here and because mixed feeding drongo parties are (at least almost completely) missing here. Stuart and Jensen (1985) report it to be common from 300 m in the eastern foothills up to at least 1400 m on the eastern slopes.

Red-faced Crimsonwing Cryptospiza reichenovii (M, P, T, G): Very common at all stations as shown by the mistnetting data (this understorey species is extremely skulking and quiet and it is really necessary to mistnet to estimate its density).

Abyssinian Crimsonwing Cryptospiza salvadorii (G): Single individuals were seen at Ukwama (1430 m, outside forest) and Tchenzema (1600 m, outside forest). These records are the first from the Ulugurus since 1922 at which time a single female was collected near Mbeta (probably =Mgeta) and later identified to species (Friedmann and Loveridge 1937)[2]. Male and female have a similar plumage in this species which in its appearance is quite similar to the female of C. reichenovii. However, at both localities (Ukwama, Tchenzema) the birds were seen under good conditions, the Tchenzema bird being observed for 5 minutes at only 10 m distance.

Lesser Seed-cracker Pyrenestes minor (G): During this survey the species was only recorded in the Kimboza Forest, where it was heard only a few times. There are several earlier records from the foothills up to 900 m, where it is an uncommon species of undergrowth (Stuart and Jensen 1985). There is also a specimen from Bagilo (Friedmann and Loveridge 1937). Generally a scarce species of undergrowth and secondary scrub of the forest edge, streamside thicket and nearby cultivation in Eastern Tanzania (Britton 1980). The species reaches its nortwestern limit in the Ulugurus (Britton 1980, Stuart and Jensen 1985)

Peter’s Twinspot (Red-throated Twinspot) Hypargos niveoguttatus (G): On this survey only recorded in the forest strip at Kigurunyembe where it was seen and one was mistnetted, and outside forest at Singiza Mission, 440 m. According to Stuart and Jensen (1985) it is common in the eastern foothill forests with a single record from Morningside, 1200 m, from the undergrowth of a plantation, probably not being a forest species at intermediate elevations. According to Britton (1980) the race H. n. macrospilotus is a widespread and reasonably common resident frequenting forest undergrowth, moist bush land and woodland thicket, thick herbage by water and dense secondary scrub, ranging throughout most of Tanzania.

Green-backed Twinspot (Green Twinspot) Mandingoa nitidula (M, G): A single individual was mistnetted in Kimboza Forest. This is the first record from that locality. Earlier Uluguru records are from Bagilo (Friedmann 1928), Morningside, 1200 m (Moyer and Baker, pers. comm. to Stuart and Jensen 1985) and "Uluguru," 1200-1500 m (Andersen in Stuart and Jensen 1985). M. nitidula is very difficult to detect except with mistnetting, and it may be not uncommon in the ground-water forests along the base of the mountains.

Oriole Finch Linurgus olivaceus (M, P, G): Seen now and then at higher altitudes during the survey. Appeared to occur at low densities. Mistnetted at Kimhandu-3 (1940 m) and Lanzi-3 (2110 m) and recorded on plots at Kimhandu-5 (2520 m) and Lanzi-2 (1920 m). Stuart and Jensen (1985) describe the species as being generally uncommon in the Ulugurus, although locally numerous where there is an abundance of seeding plants.

Appendix 5.6. Observations of non-forest bird species from our survey.

The records of “non-forest bird species” (made inside and outside forest) from our survey are listed below. Only little emphasis was laid on “birding” outside forest. Earlier records are not mentioned below unless we have found it relevant, we refer to Stuart and Jensen (1985). We follow Stuart and Jensen's practice by including only species which have been observed above 900 m in the Ulugurus and have not included observations from Morogoro Town[3]. A few species recorded in Kimboza F.R. but not known from above 900 m are included, however. For several of the species there were no earlier records published specifically from the Ulugurus. Very common and widespread species may have been “ignored” by earlier fieldworkers to some extent. For other species the lack of records may best be explained by a shortage of fieldwork.

Double underlined: Species seen inside forest during our survey and which are prominent members of the forest community on at least some of the localities mentioned. Single underlined: Other species seen inside forest during our survey (excluding most species seen over the forest except for a few swift and raptor species).

Hamerkop Scopus umbretta: Singiza 440 m, Ukwama 1430 m, Lanzi 900 m, Bunduki c. 1300 m, Kigurunyembe 700 m (where seen several times inside the forest along a stream). African Black Duck Anas sparsa: Seen swimming in the Mgeta Stream above Lanzi (1860 m). Earlier Uluguru records are from 600-1200 m. Honey Buzzard Pernis apivorus: Singiza 440 m, Ukwama 1400 m. First published record from the Ulugurus of this Palearctic migrant. Bat Hawk Machaeiramphus alcinus: A single individual was seen in Kimboza Forest at 300 m. First published record from the Ulugurus. Palm-nut Vulture Gypohierax angolensis: Kimboza 300 m. Earlier records are from the “the foothills of the Ulugurus” (Britton 1980) and from Kimboza Forest (Stuart and Jensen 1985). Gymnogene (African Harrier Hawk) Polyboroides typus: Kimhandu 1430 and 2030 m, Lanzi 2030-2100 m, Tegetero 1300-1320 m. Seen hunting well inside forest at all these three localities. Also seen far from forest near Singiza village at 440 m and close to forest in Ukwama village at 1430 m. Earlier records are from c. 300 m (eastern foothill forests) and 2140 m (Lupanga Peak). Gabar Goshawk Melierax gabar: Tchenzema 1700 m where it nested in a big Ficus tree. First published record from the Ulugurus. Little Sparrowhawk Accipiter minullus: One individual was seen inside forest at Lanzi, 2000 m. Earlier records are from between 300 and 1200 m. In Tanzania generally a species of woodland but occurs at low densities in many forests. Common Buzzard Buteo buteo (Steppe Buzzard Buteo (b.) vulpinus): Ukwama 1400-1450, Lanzi 2030 m, Tchenzema 1700 m. Palearctic migrant. Steppe Eagle Aquila nipalensis: Lanzi 1680 m. First published record from the Ulugurus. Wahlberg's Eagle A. wahlbergi: Lanzi 900 m, Kimboza 300 m. First published record from the Ulugurus. Booted Eagle (African Hawk Eagle) Hieraaetus spilogaster: One was seen at 1400 m in the Kimhandu area. Long-crested Eagle Lophaetus occipitalis: Kimhandu 1500-1990 m. Occasionally seen soaring over the forest and open country side in the Kimhandu area. May breed in the forest edges. Lesser Kestrel Falco naumanni: Bunduki 1000 m. First published record from the Ulugurus. Lanner Falcon F. biarmicus: Ukwama 1500 m. First published record of this huge falcon from the Ulugurus. Peregrine Falcon F. peregrinus: Bunduki 1700 m (breeding record). Red-eyed Dove Streptopelia semitorquata: Kigurunyembe at 700-1000 m in woodland and at 650-850 m in the ecotone between forest and adjacent woodland. Emerald-spotted Wood Dove Turtur chalcospilos: Kimboza 300 m. Blue-spotted Wood Dove T. afer: Tchenzema 1600 m. Green Pigeon Treron calva: Seen frequently in Kimboza Forest during the visit by JF and JK. Near Bunduki a few individuals were seen and in the foothills below Bunduki large flocks were seen in the miombo fire-climax zone. At Kigurunyembe it was heard at c. 700 m in the forest strip and seen in woodland just outside the strip. There were no earlier published records from the Ulugurus apart from an observation mentioned in Friedmann and Stager (1964) without any information on altitude and locality (Stuart and Jensen 1985). Stuart and Jensen (1985) did, however, expect the species to be present. It inhabits a broad range of habitats (Britton 1980, Urban et al. 1986), being mostly confined to edges and clearings when occurring in forest. Fischer’s Lovebird Agapornis fischeri: One pair (breeding?) was seen at 600 m near buildings at Teachers College, Kigurunyembe, on 7 November and again 7-8 December. Britton (1980) did not mention any records from the Ulugurus. This species is treated as Near-threatened in Collar et al. (1994). Red-chested Cuckoo Cuculus solitarius: Singiza 440 m, Ukwama 1400-1430 m (common), Kimhandu 1520-1600 m, Lanzi 900 m (fairly common). A few individuals were recorded inside forest at Kimhandu though never very far from the forest border. Klaas's Cuckoo Chrysococcyx klaas: A few individuals were heard singing inside the forest at Kimhandu 1550-1940 m, Tegetero 1345-1600 m and Kimboza 300 m. At Kigurunyembe it occurred in the forest edge at 650-850 m and in woodland outside the forest strip. Also recorded at Bunduki 1300 m and near Singiza village at 440 m. There were several earlier records from forest and non-forest habitats in the Ulugurus, from the foothills up to 1800 m. Britton (1980) describes the species as widespread and reasonably common in woodland, forest edges and clearings, thickets and gardens from sea-level up to 3000 m. White-browed Coucal Centropus superciliosus: Ukwama 1430-1500 m, Lanzi 900 m, Kimboza 300 m (common), Kigurunyembe in woodland outside the forest strip (common). Bat-like Spinetail (Böhm's Spinetail) Neafrapus boehmi: Kimboza 300 m. Stuart and Jensen (1985) report this species to be fairly common in Kimboza Forest. According to Fry et al. (1988) this species inhabits airspace around trees in a variety of woodlands and also clearings and edges of evergreen lowland forest. African Palm Swift Cypsiurus parvus: Kimhandu 1940 m (common), Kimboza 300 m (very common). Mottled Swift Tachymarptis aequatorialis: Ukwama 1430-1520 m, Kimhandu 2145 m. One earlier published record from the Ulugurus (from Lupanga Peak). According to the map of distribution shown in Fry et al. (1988) this species (which occurs in rocky crags in highlands up to 3000 m) is in Tanzania restricted to highlands in the northern and south-western part of the country with a few scattered records elsewhere. [[Nyanza Swift A. niansae: Kimhandu 1600 m, Lanzi 2130 m. First published records from the Ulugurus. According to Britton (1980) and Fry et al. (1988) this partially migratory swift has its southernmost population in Arusha N.P. in northern Tanzania. Identification of swifts is sometimes tricky, so it is best to regard our two single observations with some caution, pending further fieldwork in the Ulugurus]]. African Black Swift A. barbatus: Kimhandu 1600 m (common), Lanzi 1900-2570 m, Tegetero 1345 m. First published records from the Ulugurus. According to Fry et al. (1988) there are only few published records from E and S Tanzania. Little Swift A. affinis: Kigurunyembe 700-1000 m, Singiza 440 m (very common and breeding). First published records from the Ulugurus of this widespread species. African White-rumped Swift A. caffer: Lanzi 900 m, Tegetero 1660 m. Speckled Mousebird Colius striatus: Morningside below 1500 m (common), Bunduki below 1500 m (common). European Bee-eater Merops apiaster: Lanzi 1920-2200 m where flocks of what we believe were migrating birds were frequently recorded, sometimes only heard soaring above the forest. Also at Kigurunyembe where seen above woodland at 700-1000 m. First published records from the Ulugurus. Broad-billed Roller Eurystomus glaucurus: Recorded fairly frequently in Kimboza Forest. Also seen at 600 m in the park area at Teachers College near Kigurunyembe. First published record from the Ulugurus. Britton (1980) describes the species as being locally common in woodland, forest edges and clearings, and open country with tall trees near water, usually below 2000 m. It is an intra-African migrant and is probably most abundant in the hot season. Green Wood Hoopoe (Red-billed Wood-Hoopoe) Phoeniculus purpureus: A single bird was seen in the lowest forest part in the Kimhandu area at 1480 m. First published Uluguru record from outside Kimboza Forest where seen several times during the visit by JF and JK and where also seen by Stuart and Jensen (1985). According to Fry et al. (1988), P. purpureus is "absent from arid zones and forest". However, during fieldwork in the Udzungwas 1991-92 (Ndundulu Forest, 1350-2400 m) it was found deep inside forest at 1500-1800 m, although in low numbers (LAH and JOS). In Mwanihana Forest on the eastern scarp of the Udzungwas it occurs from 1200 to 1800 m (Stuart et al. 1987). Further fieldwork in the Uluguru montane forests might reveal a similar occurrence inside forest here. According to N. Burgess (in litt.) it also occurs in coastal forests. Britton (1980) describes it as a widespread resident and wanderer in woodland and other habitats with tall trees up to 2800 m. Spot-flanked Barbet Tricholaema lachrymosa: Bunduki 1300 m. First published record from the Ulugurus. Greater Honeyguide Indicator indicator: Kimboza 300 m. First published record from the Ulugurus. Black Saw-wing Psalidoprocne holomelas: Ukwama 1400-1600 m, Kimhandu 1520-2520 m (common), Lanzi 1840-2030 m (common), Tegetero 1300 and 1680 m (common), Tchenzema 2100 m, Kigurunyembe 650-850 m (common). Heard often above the forest. Sometimes forages low over the canopy and at a few occasions even below the canopy along streams. Seen frequently near the forest edges. Red-rumped Swallow Hirundo daurica: Ukwama 1400-1520 m, Kimhandu 1520-2520 m (common to very common), Lanzi 1680-2570 m (common), Tegetero 1345-1770 m (common), Tchenzema 1700 m. In many places seen near the forest edge. Flocks of up to 150 individuals were seen at Kimhandu-5 (2520 m). African Rock Martin H. fuligula: Singiza 440 m, Ukwama 1400-1430 m (common), Lanzi 1680 m (common). Wire-tailed Swallow H. smithii: Singiza 440 m (very common, nesting), Kimboza 300 m. European Swallow H. rustica: Ukwama 1400 m (common), Lanzi 2500-2570 m (common to very common). Grey Wagtail Motacilla cinerea: Lanzi 1850-1890 m where 1-2 birds were seen on more occasions in the Mgeta stream (probably same bird(s) seen each time), sometimes where the river passed a meadow but also sometimes inside forest. First published record from the Ulugurus of this Palearctic migrant. Mountain Wagtail M. clara: Ukwama 1430 m, Kimhandu 1520-1710 m, Lanzi 1900 m, Tegetero 1480-1720 m, Bunduki 1500 m, Kigurunyembe 700 m (heard once). Seen in mountain streams both inside and outside forested areas. African Pied Wagtail M. aguimp: Singiza 440 m. Common Bulbul Pycnonotus barbatus: Singiza 440 m, Ukwama 1430-1540 m (very common), Kimhandu 1520-1550 m, Lanzi 900 m (very common), Tchenzema 1700 m, Kimboza 300 m (common), Bunduki up to 1500 m (common), Kigurunyembe 600-850 m. Recorded most frequently outside forest but was also seen and heard in edge habitats, sometimes extending a little into forest. At Kigurunyembe it was fairly common even inside the forest. Cape Robin Cossypha caffra: Ukwama 1400-1600 m (common), Tchenzema 1500-1700 m, the Lukwangule Plateau 2500 m, Morningside 1500 m. Heuglin’s Robin (White-browed Robin-Chat) C. heuglini: Lanzi 900 m (common), Morningside 1500 m (common) and at Mgeta and Tegetero. Eastern Bearded Scrub Robin Cercotrichas quadrivirgata: A character bird in the forest strip at Kigurunyembe up to at least 850 m, one was mistnetted at this locality (Table 5.12). According to Britton (1980) this species is a common resident of bush land and woodland thickets, forest undergrowth and gardens from sea-level to 1000 m, exceptionally as high as 1800 m in the North Pare Mountains. Stuart and Jensen (1985) did not mention the species (possibly regarding it a non-forest species with no records from above 900 m?). Stonechat Saxicola torquata: Ukwama 1400-1500 m, Kimhandu 2520 m (1-2 pairs seen on the high altitude meadow), Lanzi 1680 m (common), Morningside 1500 m (common), Bunduki 1000 m (common). Seen in many places near villages and forest edges. Whinchat S. rubetra: Ukwama 1430-1600 m (common). First published record from the Ulugurus of this Palearctic migrant. Cinnamon Bracken Warbler Bradypterus cinnamomeus: Kimhandu 2145-2600 m, Lanzi 1710-2500 m, Tchenzema 1700-2500 m. Surprisingly there were only two earlier published records from the Ulugurus of this species (Stuart and Jensen 1985). We found the species to be very common around our Kimhandu-5 (2520 m) camp on the big meadow and in the surrounding elfin forest (especially near the edge), which is clearly reflected in the plot data (Figure 5.1). At this locality it is found sympatrically with its congener B. mariae (see Appendix 5.5), although being more confined to edge habitats than the latter. We also recorded B. cinnamomeus frequently on the Lukwangule Plateau on the visits to this area above Tchenzema and Lanzi. Single individuals were recorded occasionally in big forest glades with bracken in the Kimhandu, Lanzi and Tchenzema areas down to 2140 m. Also recorded on a meadow surrounding the Mgeta stream at 1860 m and near the lower forest edge at Lanzi. Britton (1980) describes it as a wide-ranging and common bird of moist thickets, giant heather, bracken-briar and the undergrowth of forest and bamboo in the highlands at 900-2400 m (mainly above 2000 m). Mountain Yellow Warbler (Mountain Flycatcher-Warbler) Chloropeta similis: Ukwama 1430 m. There are two earlier published records from the Ulugurus (one from Nyingwa and one from Bunduki) (Friedmann 1928, Baker pers. comm. to Stuart and Jensen 1985). According to Britton (1980) this species occurs in shady places in bracken, bamboo and forest edges at 1800-3400 m. Willow Warbler Phylloscopus trochilus: Lukwangule Plateau 2500 m. First published record from the Ulugurus of this Palearctic migrant. Trilling Cisticola Cisticola woosnami: Singiza to Nyamigadu (very common) 1000-1500 m, Ukwama 1430 m (common), Lanzi 1680 m (fairly common), Tchenzema 1700 m (common), Morningside below 1400 m. Tawny-flanked Prinia Prinia subflava: Kimhandu 900 m (common), Bunduki common around 1000 m, Kigurunyembe 700-1000 m (common). Recorded often during the walk from Singiza to Nyamigadu (altitude not noted). Slaty Flycatcher Melaenornis chocolatinus (White-eyed Slaty Flycatcher Melaenornis fischeri): Ukwama 1430 m, Tchenzema 1700 m. First published records from the southern section of the Ulugurus. Britton (1980) describes the species as common in forest edges, woodland and gardens at 1300 to 3000 m. Black-and-white Flycatcher Bias musicus: Seen in Kimboza F.R. and just outside forest at Bunduki (1300 m). Also outside forest near Bwakira Juu (c. 340 m) and Dakawa (c. 300 m) villages on 30 September. Britton (1980) describes it as an adaptable species which is common in tall trees in gardens, woodland, bush land and forest edges as well as in the forest canopy (ranging from sea-level to 1700 m). One of the species which it is difficulty to categorize into forest/non-forest species. Arrow-marked Babbler Turdoides jardineii: Kigurunyembe at 700-1000 m in woodland. Spotted Creeper Salpornis spilonotus: Ukwama 1430 m on a tree trunk in a field. First published record from the Ulugurus. Collared Sunbird Anthreptes collaris: This species was recorded only in Kimboza Forest (fairly common) and in the forest strip at Kigurunyembe 650-850 m (few seen and one mistnetted). Stuart and Jensen (1985) describe the species as being much commoner at low altitudes where it is more of a forest bird but there are records from up to 1800 m (forest and non-forest habitats). During a survey in the Udzungwa Mountains in 1991-92 (Ndundulu Forest, 1350-2400 m) A. collaris was seen deep inside forest up to at least 1550 m although usually at low densities (LAH and JOS). It therefore surprised us that we did not record it at any of the localities in the Uluguru North and South F.R.s. One of the species which it is difficulty to categorize into forest/non-forest species. Variable Sunbird (Yellow-bellied Sunbird) Nectarinia venusta: Tchenzema 1600 m, Morningside 1500 m, Bunduki 1000 m (common), Kimboza 300 m, Kigurunyembe 700-1000 m (woodland). Also recorded above Singiza (altitude not noted) and in other cultivated land areas passed on our way to and from the forests. Southern Puffback (Black-backed Puffback) Dryoscopus cubla: The species was recorded frequently in Kimboza F.R. during the visit by JF and JK. Also Stuart and Jensen (1985) describes it as being common in Kimboza. It was fairly common in the forest strip at Kigurunyembe at 650-850 m and was also recorded in woodland outside this strip. Also at Tchenzema we found it inside forest (recorded on two of 25 plots, see Table 5.9) whereas it was not recorded on the other forest localities visited. Recorded at the following non-forest localities during our survey: Singiza 440 m (common), Nyamigadu area 1500 m, Lanzi 900 m (fairly common), Tchenzema 1700 m, Bwakira 340 m, and Dakawa 300 m. Apart from Kimboza the only earlier published records were from Bunduki and Morningside. Britton (1980) describes the species as wide-ranging and often common up to 2200 m in forest edges and strips, woodland, gardens and sometimes the interior of larger tracts of forest. It would perhaps be most correct to regard it a “forest species”. Tropical Boubou Laniarius aethiopicus: Recorded frequently in Kimboza F.R. during the visit there by JF and JK. Above Tchenzema it was recorded inside forest on four of 25 plots (see Tables 5.9; the four plots were at 2000-2050 m). Records made at non-forest localities during our survey: Singiza 440-800 m (very common), Ukwama 1500 m, Lanzi 900 m (fairly common), Tchenzema 1700 m and Morningside below 1500 m. Black-headed Tchagra (Black-crowned Tchagra) Tchagra senegala: Singiza 440 m, Bunduki below 1300 m. Only the second and third published records from the Ulugurus since Friedmann and Loveridge's (1937) first record. Orange-breasted Bush Shrike (Sulphur-breasted Bush Shrike) Telophorus sulfureopectus: Bunduki 1300 m. First published record from the Ulugurus. Retz's Red-billed Helmet Shrike Prionops retzii: Stuart and Jensen (1985) describe this species as common in Kimboza and it was also recorded there on some occasions during our survey. There is an earlier record from Morningside, 1200 m. According to Britton (1980) the species is widespread, but seldom common, in woodland from sea-level to 1990 m (mainly below 1500 m). Fork-tailed Drongo Dicrurus adsimilis: Bunduki 1000 m (common). First published record from the Ulugurus of this common species. Pied Crow Corvus albus: Kimhandu 1940 m, Lanzi 900 m (common). White-necked Raven C. albicollis: Ukwama 1430, Kimhandu 1520-2520 m (common), Lanzi 900-2500 m, Tegetero 1345-1950 m, Tchenzema 2500 m, Bunduki 1300 m (breeding record), Kigurunyembe 700-1000. A common species which was often seen and heard soaring above the canopy or surrounding cultivations. Red-winged Starling Onychognathus morio: At Kimhandu (the only locality where we recorded it on this survey) it was seen occasionally in the canopy and more often in big flocks flying over the forest at 1700-2000 m. Stuart and Jensen (1985) describe it as widely distributed from the foothills up to at least 1800 m, especially in the vicinity of rock faces and settlements. Bertram's Weaver Ploceus bertrandi: Tchenzema 1600 m. Several seen from the roadside west of the mountains. [[Speke's Weaver P. spekei: Ukwama 1430 m. There were no earlier published records from the Ulugurus. According to Britton (1980), this species (which is a gregarious resident and wanderer, frequenting a variety of open country with trees or bushes at 1200-2000 m, especially in modified habitats near habitation) is not known from south of the Arusha-Kilimanjaro area. Recently it has been found further south in Singida, however (N.E. Baker in litt.). The species looks relatively characteristic on our colour plates, but since plumage variation in relatively similar-looking species, e.g. African Masked Weaver (Vitelline Masked Weaver) P. velatus cannot be excluded (the eye-colour also differing between the two species, however), it is best to regard our single observation with some caution, pending further studies in the Ulugurus]]. Golden Weaver P. subaureus: Lanzi 900 m (common). Not recorded since the specimens collected by Andersen (Andersen in Stuart and Jensen 1985) and by Loveridge (Friedmann 1928). African Masked Weaver (Vitelline Masked Weaver) P. velatus: Ukwama 1430 m. First published record from the Ulugurus. African Firefinch (Blue-billed Firefinch) Lagonostica rubricata: Morningside 1500 m. Swee Waxbill (Yellow-bellied Waxbill) Estrilda quartinia: Ukwama 1400-1600 m (common), Lanzi 900 m (common to very common), Tchenzema 1700 m (very common), Morningside below 1500 m (common). Common Waxbill E. astrild: Ukwama 1400-1450 m (very common), Tchenzema 1700 m. Bronze Mannikin Lonchura cucullata: Singiza 440 m (very common). Red-backed Mannikin L. nigriceps: Singiza 440 m. In Britton (1980) called Rufous-backed Mannikin L. bicolor nigriceps, a subspecies of the Black and White Mannikin L. bicolor. African Citril Serinus citrinelloides: Tchenzema 1600 m. Cabanis's Bunting Emberiza cabanisi: Ukwama 1430 m.

Appendix 5.7. Observations of mixed feeding parties from our survey.

Many of the species occurring in the Eastern Arc forests take part in mixed feeding parties regularly, and a considerable diversity of flock structures have been seen in Eastern Arc Mountains, from very mobile large parties which go on for hours with Dicrurus ludwigii and other core species ("drongo parties") to brief activity bursts involving few species, none of which can be called "core species" (JF, LAH and JOS). Understorey species sometimes form their own parties.

Our survey revealed some interesting differences between the submontane areas of the Uluguru North F.R. on one hand and the montane areas of the Uluguru South and North F.R.s on the other. Only few parties were studied in detail, however, and the descriptions below are therefore kept in general terms.

Tegetero. Mixed feeding drongo parties (here defined as parties including typical party species like Dicrurus ludwigii, Ploceus bicolor, Phyllastrephus flavostriatus, Apalis melanocephala, Telophorus nigrifrons and Coracina caesia) were a very typical phenomenon below 1450 m.

Other species seen in the drongo parties include (* indicates that most observations below 1500 m are of birds in drongo parties): Apaloderma vittatum, Stactolaema olivacea*, Pogoniulus leucomystax, Dendropicos griseocephalus*, Terpsiphone viridis*, Andropadus olivaceiceps, A. masukuensis, Apalis chariessa*, A. thoracica, A. chapini, Phylloscopus ruficapillus, Trochocercus albonotatus, Muscicapa adusta, Laniarius fuelleborni, Nectarinia loveridgei, N. olivacea, Onychognathus walleri and Zosterops senegalensis.

The drongo parties were generally immobile, as they often stayed in almost the same area for hours and sometimes foraged quietly for long periods. It was our impression that the density of parties was high, but we often found it difficult to distinguish different parties with certainty. We believe that the low mobility and the frequent silence of the parties should be explained by the fact that we were in the beginning of the breeding season (see Appendix 5.11). Some structurally loose parties without Dicrurus ludwigii but including both Phyllastrephus flavostriatus and Ploceus bicolor, sometimes P. bicolor, Coracina caesia and Telophorus nigrifrons but not Phyllastrephus flavostriatus, were observed between 1460 and 1550 m. A few parties without any of the core species of drongo parties were seen above 1700 m. A few parties of understorey species were seen at Tegetero.

Kimhandu and Lanzi. Whereas mixed feeding parties were common at Tegetero near the two lowest stations, only few and brief mixed feeding parties were recorded at Kimhandu and Lanzi where the forest is generally situated above 1500 m. We did not record the typical feeding party species Ploceus bicolor, Phyllastrephus flavostriatus and Apalis melanocephala in the southern section, and Dicrurus ludwigii was recorded on only a single occasion (a pair seen in the lowest part of the forest near Kimhandu-6 [1540 m], not in a feeding party; first published record from the Uluguru South). Two other species often seen in drongo parties but which generally have a wider altitudinal tolerance were present: Telophorus nigrifrons and Coracina caesia. At Kimhandu T. nigrifrons occurred at unexpectedly low densities and C. caesia was seen only near our 1940 m station (few observations). At Lanzi, these two species appeared more common than at Kimhandu but clearly occurred at much lower densities than near the two lowest stations at Tegetero.

The following species were seen participating in feeding parties or brief feeding bursts: Pogoniulus leucomystax, Alcippe abyssinica, Andropadus olivaceiceps, A. masukuensis, A. virens, Phyllastrephus placidus, Pogonocichla stellata, Apalis thoracica, A. chapini, Phylloscopus ruficapillus, Trochocercus albonotatus, Batis mixta, Terpsiphone viridis, Laniarius fuelleborni, Telophorus nigrifrons, Muscicapa adusta, Nectarinia loveridgei and Zosterops senegalensis.

Species that frequently took part in drongo parties at Tegetero were generally seen feeding individually at Kimhandu and Lanzi, e.g. Terpsiphone viridis. It should be noted that we visited the Kimhandu and Lanzi areas in the beginning of the breeding season (see Appendix 5.11) and that in general mixed species flocks are a more common phenomenon in the dry season.

Tchenzema. No feeding parties were recorded during the brief visit here.

Kimboza. Many mixed feeding flocks were seen with Dicrurus ludwigii almost invariably present. Like at Tegetero the parties appeared to stay long periods in a small area. It is of interest that Andropadus virens, which foraged solitarily and within its territories in the lower parts of the Kimhandu area, was generally very active in mixed feeding parties in Kimboza F.R. This, together with the above-mentioned example with Terpsiphone viridis, suggests that Dicrurus ludwigii (and other core species) are important for the formation of mixed feeding parties.

Comparing with the Udzungwas, it is interesting that during research in the Ndundulu Forest in the Udzungwas 1991-92 (LAH and JOS) mixed feeding parties of the drongo party type were a common phenomenon up to c. 1800 m.

Appendix 5.8. Seasonal altitudinal migration - indications from this survey and from the literature.

(See Footnote 4 in the end of this appendix for an explanation of the phenomenon of seasonal altitudinal migration). Some of the bird species that breed in the mountain forests appear to depend on low altitude forest in the dry season. For the conservation of the population of such species it is necessary to protect not only their mountain habitats, the lowlands must also be protected.

Two of the restricted-range species in the Ulugurus can be found in the foothill forests at least part of the year, one of them (A. fuelleborni) having been found in good densities. In this section we have collected those of our own observations that are related to the phenomenon of vertical migration and have supplemented them with previous records of the same kind (listed in Stuart and Jensen 1985), in an attempt to assess what is actually known about this phenomenon in the Ulugurus at present. For some of the species the records could be explained by other types of movements than "ordinary" seasonal vertical migration (see footnote) or the species may prove to breed in the lowland forests. It is clear that the knowledge on vertical migration in the Ulugurus is still scanty.

1. Aplopelia larvata. Kimboza Forest, 300 m, in July (Stuart and Jensen 1981). One heard singing in the forest strip at Kigurunyembe, 700 m, on 7 December 1994 (JF and JK) - unusual with an observation from such a low altitude at this time of the year.

2. Cercococcyx montanus. One juvenile male from "East Uluguru", 400 m, in February (Andersen in Stuart and Jensen 1985). Recorded in Kimboza Forest in October during our survey. So, the species can be found in the lowlands also in what is normally considered the early breeding season of mountain forest birds.

3. Apaloderma vittatum. Kimboza Forest in August (Andersen in Stuart and Jensen 1985) and Kibungo Forest, 300 m, in June and September (Moreau in Stuart and Jensen 1985; Fuggles-Couchman, pers. comm. to Stuart and Jensen 1985). Andersen collected a specimen at 400 m "east of Uluguru" (Britton 1981).

4. Andropadus olivaceiceps. Andersen specimen from Kimboza Forest in July (Stuart and Jensen 1985). Kibungo Forest in June (Moreau in Stuart and Jensen 1985).

5. Zoothera gurneyi. Single bird heard at Kigurunyembe, 700 m, on 7 December 1994 (JF and JK) (mentioned here though admittedly not in the foothills).

6. Alethe fuelleborni. Kibungo and Kimboza Forests, 250-300 m, in June and July (Moreau in Stuart and Jensen 1985; Stuart and Jensen 1981). Stuart and Jensen (1981) described it as being common in Kimboza Forest in July 1981. Two individuals were mistnetted in Kimboza Forest on 17-19 October during our survey (of a total capture of 19 birds at that locality!). These captures are from what is normally considered the early breeding season of mountain breeding species. Further fieldwork in Kimboza is necessary to clarity the significance of this forest for A. fuelleborni. JF has found the species to be common in September-October in many other Tanzanian foothill forest (Rubehos, Mt. Kanga, Ngurus).

7. Pogonocichla stellata. All records from below 1000 m are from June and July, including several from the Ngerengere River, 600 m, 10 km N of the Uluguru Mountains (Andersen in Stuart and Jensen 1985). Recorded frequently during fieldwork in Kimboza Forest July 1981 (Stuart and Jensen 1981). Heard once at Kigurunyembe, 700 m, on 7-8 December 1994 (JF and JK).

8. Bradypterus mariae. Kibungo Forest in June (Moreau in Stuart and Jensen 1985). Kimboza Forest medio October 1994, single bird (JF and JK).

9. Trochocercus albonotatus. Kibungo Forest in June (Moreau in Stuart and Jensen 1985).

10. Modulatrix stictigula. There is an Andersen record from bushland at 600 m (Britton 1981, Stuart and Jensen 1985). This record is extraordinary as there are no other indications of migratory/exploratory movements of this species.

11. Telophorus nigrifrons. Kibungo and Kimboza Forests in June and July (Moreau in Stuart and Jensen 1985; Stuart and Jensen 1981).

12. Poeoptera kenricki. Kimboza Forest in July (Stuart and Jensen 1981).

Also Terpsiphone viridis, Smithornis capensis, Coracina caesia and Muscicapa adusta must be suspected to migrate altitudinally (see footnote). They are not included in the list above because they cannot be considered strictly montane, both being known from coastal lowland forests too. Terpsiphone viridis and Muscicapa adusta had not previously been recorded in Kimboza F.R. but were recorded there during our visit.

When we arrived to the Uluguru South F.R. on 1 October, all local and intra-african migrants species for which we assume the seasonal fluctuations are conspicuous (e.g. Pogonocichla stellata and Terpsiphone viridis) were present and common. We did not see any signs of an increasing density during our survey (due to frequent movings between camps we could have overlooked day to day changes, however). This indicates that most of the populations had arrived by the time we started the survey (this might be different from other Eastern Arc Mountains e.g. montane and submontane zones of the Udzungwas [pers. comm., Louis A. Hansen]).

The distance to fly from the nearest points of the Uluguru North and South F.R.s to groundwater forests in the foothills is at least 15 km. The distance from e.g. the Kimhandu area is much longer (at least 25 km). Scharff et al. (1982) mentioned that the migration pattern may differ between the eastern and the western slope of the Uluguru North F.R.

[4]

Appendix 5.9. Further notes on the general characteristics of the forest avifaunas at the localities visited.

5.9.a. Submontane evergreen forest.

In addition, for three of the species assumed to exhibit seasonal fluctuations in abundance (see Appendix 5.8) the submontane habitat supports much higher densities than the montane. These are Smithornis capensis, Terpsiphone viridis and Cercococcyx montanus. Smithornis capensis was very uncommon in the Uluguru South: At Kimhandu it was recorded only at 1700 m altitude and at Lanzi it was not recorded at all. This species clearly prefers submontane forest. Cercococcyx montanus also occurred at low densities in the Uluguru South F.R., being more common in the submontane and lower montane areas in the Uluguru North F.R. For a further discussion of these two species: see Appendix 5.5. Terpsiphone viridis was relatively common in the montane forests in the Uluguru South F.R. but clearly occurred at higher densities in the submontane areas of the Uluguru North.

Tegetero-1 (1345 m, in the submontane belt) and Tegetero-2 (1535 m, higher submontane-lower montane belt) are the most species rich of our study areas (Table 5.5[5]). The high numbers should be explained by the occurrence of typical drongo party species plus species recorded nowhere else (e.g. Malaconotus alius and Apalis chariessa) in combination with the occurrence of most of the species occurring in the montane zone. The species richness is reflected in the plot data but not in the mistnetting data - the set of species prone to mistnetting is largely the same as on many other stations.

5.9.b. Montane evergreen forest.

Some differences caused by variations in the character of the forest and the altitudinal position of the lower forest edge: The avifauna appears to be of a montane character further down in the study area at Lanzi than in the study area at Kimhandu, since three species typical of high altitudes (Andropadus neumanni, Cossypha anomala and Apalis thoracica) are common down to the lower forest edge at Lanzi whereas they are not common below 1900 m at Kimhandu (Tables 5.8 and 5.9). For Alethe fuelleborni, a restricted-range species which at Kimhandu was most common at the lower stations, the lack of forest below 1600 m at Lanzi, probably in combination with the apparently more montane character at Lanzi, appears to have caused lower densities here - only 1.6 individuals were mistnetted per 2500 MNH (Table 5.8).

At Tchenzema on the drier western slope of the mountains, the sharp lower forest edge is at 1900-2050 m. The forest interior is strongly disturbed by man in the lower part of the gradient at this locality. The tree height was generally only 10-15 m, even in the lowest parts. The open forest structure (combined with proximity to lower edge) at Tchenzema is probably the explanation of the occurrence of two non-forest species Dryoscopus cubla[6] (scored on two of 25 plots; 2000 m, 2150 m) and Laniarius aethiopicus[7] (scored on four of 25 plots; 2000 m, 2000 m, 2000 m, 2050 m). (Interestingly, neither L. aethiopicus nor Dryoscopus cubla were recorded by Scharff et al. (1982) during their brief visit to the Tchenzema area in 1981). These two species were not recorded at Kimhandu, Lanzi or Tegetero. The open structure of the forest at Tchenzema also favours species like Laniarius fuelleborni and Cossypha anomala, their abundance being well illustrated by the plot data (Table 5.9, Figure 5.1.m). Both were common not only in the montane zone above Tchenzema but also on the Lukwangule Plateau.

Certain species, which could potentially be present, appear to be missing at Tchenzema. It is not surprising that the following species have not been recorded, since they are mainly submontane and lower montane species and are also absent or only recorded rarely (always below 1900 m) on the eastern slopes at Kimhandu and Lanzi:

Smithornis capensis (very few records from Kimhandu, none from Lanzi), Dicrurus ludwigii (one record from Kimhandu, none from Lanzi), Phyllastrephus flavostriatus, Alethe fuelleborni[8], Apalis melanocephala and Ploceus bicolor.

In addition it also seems certain that some species which have been recorded above 1900 m at Kimhandu and Lanzi appear to be missing or occur at very low densities at Tchenzema since they went unrecorded, however the survey was brief survey but they were neither recorded by the brief survey described in Scharff et al. (1982): Ceratogyma brevis, Sheppardia sharpei (also lacking at Lanzi-3 [2110 m], however), Terpsiphone viridis, Cercococcyx montanus (uncommon at Kimhandu and Lanzi, however) and Poeoptera kenricki (generally uncommon and therefore easy to overlook).

Species recorded at Tchenzema by Scharff et al. (1982) but not during this survey (indicating that they are scarce) include: Apalis chapini, Phyllastrephus placidus and Telophorus nigrifrons.

Andropadus masukuensis was scarce at Tchenzema, probably because of the lack of epiphytes on the tree trunks in this disturbed habitat, combined with altitudinal effects.

At Lanzi-2 (1920 m), the number of species mistnetted is low compared to the general pattern whereas at Lanzi-1 (1710 m), only 20 minutes walk from Lanzi-2, the number is high compared to the general pattern (Figure 5.2). In addition, the plot data indicate that the species diversity is higher at Lanzi-1 (34 forest species on 16 plots) than at Lanzi-2 (29 forest species on 17 plots) (Figure 5.3). On the dawn tape recordings 21 forest species were recorded at Lanzi-1 against 16 forest species at Lanzi-2. A possible explanation for the low bird species richness at Lanzi-2 is that Lanzi-1 is situated close to the forest edge in a mixture of clearings and dense forest (leading to a high diversity in the mosaic of microhabitats?) whereas Lanzi-2 is situated on a broad ridge where the vegetation is structurally more uniform and drier. There are no indications of a higher general bird density at Lanzi-1 than at Lanzi-2: the mistnet catch rate at Lanzi-2 (75.8 individuals per 2500 MNH, Table 5.8) is somewhat higher than at Lanzi-1 (59.1 individuals per 2500 MNH) and actually quite high, and the mean number of species per plot is rather similar for the two stations (Table 5.9).

5.9.c. Lowland semi-evergreen forest.

Kimboza F.R. holds a higher percentage of species which we have categorized as non-forest species than do the Uluguru North and South F.R.s. Examples of non-forest species that were recorded frequently during the brief visit by JF and JK are Green Pigeon Treron calva, Broad-billed Roller Eurystomus glaucurus, Collared Sunbird Anthreptes collaris, Dryoscopus cubla, Laniarius aethiopicus and Retz’s Red-billed Helmet Shrike Prionops retzii with records also of other non-forest species inside forest or near edges (see Appendix 5.6. Stuart and Jensen [1985] mention some additional species). Many of these lowland species admittedly occur in many East African lowland forests or in the edges of these, for which reason one could be tempted to classify them as forest species, but they are so widely distributed in non-forest habitats that we decided to classify them as non-forest species. In addition, some Kimboza species categorized in this report as forest species have a rather broad habitat range, e.g. Muscicapa caerulescens, Bias musicus, Pyrenestes minor and Hypargos niveoguttatus. When JF and JK visited the forest during our survey, the ground stratum was very dry, and most species apparently foraged in the mid and upper strata.

The significance of lowland forest for montane breeding species suspected to migrate downwards in the off-breeding season was mentioned in Appendix 5.8.

Kimboza F.R. is apparently the only of the foothill forests that is reasonably well investigated. The larger Ruvu F.R. is adjacent to Kimboza but we have not come over any ornithological records published with Ruvu mentioned as locality. Ruvu is disturbed locally from ruby mining, licensed and unlicensed, as the reserve is one of the most important localities for this gemstone in Tanzania. For tree species Lovett and Pócs (1993) mention that the western half of Ruvu F.R. is covered with seasonal lowland forest similar to that of Kimboza, but with fewer species. Ruvu may hold an important avifauna similar to that of Kimboza but surveys are necessary before we can say anything about this. In addition, the other forested parts of the karst limestone landscape below the eastern slopes of the Uluguru Mountains may hold important forest avifaunas. There are several humid thickets along the edge of the Mwina and Mtego river plains.

Appendix 5.10. Notes on Arcanator orostruthus, Swynnertonia swynnertoni and Anthreptes pallidigaster.

Dappled Mountain Robin (Dabblebreast) Arcanator orostruthus. Known from Mount Namuli in Mozambique (from two birds collected at 1465 m in the 1930ies), the Usambara Mountains (discovered in the early 1930ies, known only from 900 m altitude near Amani) and the Udzungwa Mountains (discovered 1981, recorded from 1250 to 1750 m, found in four forest patches and locally relatively common) (Collar and Stuart 1985, Moyer 1993, Dinesen et al. 1993), the three populations being classified as different subspecies. In the Ndundulu and Nyumbanitu Forests in the Udzungwa Mountains, the species appears to prefer luxuriant forest, especially where there is dense undergrowth of tall herbs (Dinesen et al. 1993, LAH and JOS pers. obs. 1994). Moyer (in litt. to Collar et al. 1994) describes its preferred habitat in the southern section of the Udzungwa Scarp F.R. as forest with closed canopy and with a dense growth of ginger Zingiberaceae in the ground stratum, between which it moves through corridors of thick growth growing around light gaps and along streams.

The forest areas below 1700 m altitude in the Uluguru North F.R. appear to offer suitable conditions for the species, e.g. in the Bagiro area where there is forest down to c. 1100 m. The species is quite vocal and relatively easy to detect (although hard to see). If overlooked in the Ulugurus, it must be very local or occur at only low densities.

Stuart (1981b) suggested that the disjunct distribution of A. orostruthus could be explained by competitive exclusion by its congener M. stictigula. However, the two species has afterwards been found coexisting at four localities in the Udzungwa Mountains, sometimes with both species occurring at good densities in areas of sympatry (Stuart et al. 1987, Jensen and Brøgger-Jensen 1992, Moyer 1993, Dinesen et al. 1993). The relatively small amount of forest habitat in the submontane and lower montane zone may be part of the explanation of its absence in the Ulugurus but this is certainly a question which calls for further studies.

Swynnerton's Robin Swynnertonia swynnertoni. Until the 1980es known only from three localities in eastern Zimbabwe (900-1700 m) and from Mount Gorongoza in Mozambique (850-1750 m) (Collar and Stuart 1985). In 1981 it was discovered in the Udzungwas where it is now known from three forest patches (1000-1700 m, being fairly common in the southern part of the Udzungwa Scarp F.R.) (Collar and Stuart 1985, Stuart et al. 1987, Jensen and Brøgger-Jensen 1992, Moyer 1993, Dinesen et al. 1993) and in the 1990’ ies, it has been discovered in several Usambara foothill forests (Evans and Anderson 1992 and 1993b, Cambridge Tanzania Rainforest Project 1994). The two populations in Tanzania form separate subspecies.

Especially the lower parts of the Uluguru North F.R. seem to offer fine conditions but, if overlooked, S. swynnertonia (who’s melancholy two-three note song is very characteristic) must occur at very low densities. Also for this species the small amount of forest habitat in the submontane and lower montane zone may be part of the explanation but the question calls for further studies.

Amani Sunbird Anthreptes pallidigaster. Known from the coastal Sokoke-Arabuko Forest in Kenya, the East Usambara Mountains (from the foothills up to 900 m) and discovered as late as 1991 in the Udzungwa Mountains (from 1400 to 1550 m, rare and occurring only locally in the Ndundulu and Nyumbanitu Forests where it is uncommon and local) (Collar and Stuart 1985, Dinesen et al. 1993, Cambridge Tanzania Rainforest Project 1994, LAH and JOS pers. obs. from the Nyumbanitu Forest 1994).

This unobtrusive and diminutive species which have recently been found in small numbers in the Udzungwas, far south of its formerly known distribution in the Usambaras and Sokoke-Arabuko, could easily be overlooked if occurring at low densities in the Uluguru North or the Kimboza/Ruvu F.R.s.

Appendix 5.11. Breeding activity noted among forest birds during our survey.

We provide evidence of breeding activity as defined below, in order to give a rough overview of the breeding status of the various species:

- Juvenile and immature birds mistnetted (listed in Tables 5.5-6).

- Birds mistnetted with brood patch (BP) 5 or with vasculated (V), wrinkled (W), scaly (S) or refeathering (RF) BP of lower score (brood patch was scored on most but not all birds mistnetted). Listed below (please compare with Table 5.5 to see total number of individuals mistnetted of the species at the various stations; species for which no individuals had BP 5 or with wrinkled or refeathering BP of lower score are not mentioned in the list below). Abbreviations for stations are those used in Table 5.5. No manual exists showing the size of the brood patch for the various species at different times of the year so it is possible that for certain species we should have listed also brood patches of lower score. Further information on, e.g. moulting patterns, brood patch scores other than those listed here etc., is available on disks.

- Nestlings and juvenile birds observed in the field (probably not all juveniles observed have been noted due to time constraints; records should be regarded as positive, not complete, evidence) (listed below).

- Observations of nesting individuals (listed below).

- Adult birds with nesting material or warning with insects in their bill (listed below).

Circaetus fasciolatus: A single juvenile bird was seen in Kimboza F.R. in October. Accipiter tachiro: At Tegetero a young bird was calling from a nest at 1345 m. Buteo oreophilus: A nestling was seen in a tree in a forest glade at Lanzi, 2000 m, 8-19 November. Stephanoaetus coronatus: We found an active nest in a big Ficus tree on the roadside in Kimboza F.R. in October. Columba arquatrix: At Kimhandu a nest with downy young was found at 2520 m on 23 of October and at 2580 m another active nest was found. Aplopelia larvata: A nest with two young birds was found at Lanzi, 2030 m, on 15 November. One with BP 5W mistnetted at Lan-3. Apaloderma vittatum: One with BP 5S mistnetted at Kim-2, one with BP 5WS mistnetted at Teg-2. Ceratogymna brevis: A male was seen feeding a female through a nesthole during the visit in Kimboza Forest on 17-19 October. Pogoniulus leucomystax: One with BP 5W at Kim-5. Dendropicos griseocephalus: Seen in a nesthole at Kimhandu, 1540 m, on 3 October and at Lanzi, 1770 m, on 5 November, but it was not clear whether the birds were breeding or just visited the nestholes for other reasons. Smithornis capensis: One with BP 5 at Teg-2. Andropadus masukuensis: One with BP 5 and one with BP 4W at Kim-2, two with BP 5W and one with BP 4W at Kim-3, one with BP 5W and one with BP 4W at Kim-4, one with BP 5W at Kim-6, one with BP 5RF at Lan-1, one with BP 5 at Lan-1, three with BP 5W at Lan-3, one with BP 5W, one with BP 5 and one with BP 4W at Teg-1, one with BP 5W and one with BP 5 at Teg-2, one with BP 5RF and one with BP 5 at Teg-3. A. neumanni: One with BP 5W at Lan-3. A. olivaceiceps: A pair warning, with insects in their bill, was seen at 1485 m altitude at Tegetero on 7 December. One with BP 4W at Kim-3. A. virens: A nest with two eggs was found close to the Kimhandu-1 (1520 m) camp on 10 October. One with BP 5 at Kim-1. Phyllastrephus placidus: One with BP 5W and one with BP 5 and one with BP 4RF at Kim-3, one with BP4W at Kim-4, one with BP 3RF at Lan-2, one with BP 5 at Teg-1, one with BP 4W at Teg-2, one with BP 5 Teg-3 and one with BP 4RF at Teg-3. Alethe fuelleborni: One with BP 5W at Teg-1, one with BP 5W at Teg-2. Sheppardia sharpei: One with BP 5VW at Teg-1, two with BP 5W at Teg-2. Cossypha anomala: One with BP 5W at Lan-1, one with BP 5W at Lan-2, one with BP 5W at Lan-3. Turdus olivaceus: Seen with nest material at Tegetero, 1610 m, on 9 December. One with BP 5W at Lan-1, one with BP 5WS at Teg-2, one with BP 5W at Teg-3. Zoothera gurneyi: One with BP 3W at Teg-2. Pogonocichla stellata: One with BP 4W at Teg-1 (48 adult individuals were scored for brood patch!). Bradypterus mariae: One with BP 4W at Lan-3. Scepomycter winifredae: One with BP 4W at Lan-1. Apalis thoracica: A family group with three young and two adults was seen foraging at close range on 16 December at 1500 m in the forest of Tegetero. One with BP 5W at Kim-5, one with BP 4W at Lan-1, one with BP 4W at Lan-2, one with BP 5W at Lan-3. Orthotomus metopias: One with BP 4W was at Kim-6, one with BP 4W at Teg-2. Batis mixta: One with BP 5 at Lan-1. Thrococercus albonotatus: One with BP 5 at Lan-1. A newly fledged young was seen at Tegetero, 1430 m, on 15 December. Alcippe abyssinica: One with BP 5W at Lan-3. Modulatrix stictigula: One with BP 4W at Kim-3, one with BP 3RF at Teg-3, one with BP 5 at Teg-1. Nectarinia olivacea: One with BP 5 and two with BP 4W at Teg-1. N. loveridgei: A newly fledged young individual was seen on 2 October at Kimhandu, 1550 m, and a female accompanied by a juvenile was observed on 12 December. One with BP 5W, one with BP 4W, two with BP 5, two with BP 5RF and one with BP 4RF at Kim-1, three with BP 5W and two with BP 5RF at Kim-2, three with BP 5W and one with BP 4W at Kim-3, four with BP 5W and one with BP 4W at Kim-4, one with BP 5W and one with BP 5 at Kim-5, three with BP 5W, four with BP 4W, one with BP 5 and one with BP 2RF at Lan-1, three with BP 5W, two with BP 4W and one with BP 5RF at Lan-2, two with BP 5RF at Lan-3, one with BP 5W and one with BP 3RF at Teg-1, one with BP 5W, two with BP 4W, two with BP 5, one with BP 5RF and one with BP 3RF at Teg-2, one with BP 5W, one with BP 4W, two with BP 5, one with BP 5RF and one with BP 4RF at Teg-3. Zosterops senegalensis: One with BP 5W and one with BP 5 at Lan-1. Laniarius fuelleborni: One with BP 5 mistnetted at Kim-5, one with BP 5 and one with BP 5W at Lan-1, one with BP 5W at Teg-1. Cryptospiza reichenovii: An individual was seen on a nest at Tegetero, 1330 m, on 6 December. One with BP 5RF and one with BP 4RF at Lan-3, one with BP 5RF at Teg-2, one with BP 5 at Teg-3. Linurgus olivaceus: One with BP 5S at Lan-3.

Juvenile birds were mistnetted only of A. masukuensis (one of 74 birds) and of Nectarinia loveridgei (five of 271 birds) (Table 5.6-7). Together with the list above of visual field observations of juvenile birds this shows that there were only few newly fledged birds when we visited the Ulugurus (although the list above is probably incomplete). Immature birds were mistnetted in more species (some immatures may have been overlooked in e.g. greenbuls though we attempted to check all individuals carefully) but this gives little information about the timing of the breeding season. Very few of the mistnetted individuals had vascularized brood patch but apart from that the amount of birds in breeding condition, determined by brood patch on mistnetted individuals, appeared to vary much between species. In the two most commonly mistnetted species, Andropadus masukuensis and Nectarinia loveridgei, we mistnetted several birds with wrinkled brood patches. In other commonly mistnetted species, e.g. Alethe fuelleborni and Pogonocichla stellata, only very few appeared to be in breeding condition. For many species the mistnet samples are too small to say anything. Most species are generally believed to start breeding at the onset of the early rains. The rains were much delayed in 1993 and did not really start before February 1994 although we experienced a few heavy rains and days with dark and moist/misty weather during the fieldwork. The delay of the rains may have led to a postponing of the breeding activities. However, no firm comparative evidence exists for this suggestion.

Appendix 5.12. Ornithology: On the methods applied and the completeness of our survey.

Below we briefly discuss the completeness of our ornithological survey in a miscellaneous way. We also comment briefly on our experience with the methods used during this survey in the hope that the notes can be of help for the planning of future surveys of African forests or for the production of a common manual for surveys. No complete manual for survey work and data analysis/presentation exists yet for Afromontane forest and it is therefore useful to leave miscellaneous notes on these issues though this report may not be the right forum. Numbers detected by the various methods are those from our three main localities only.

Mistnetting. With this method we recorded 32 forest species and one non-forest species (Bradypterus cinnamomeus) (Kimhandu, Lanzi and Tegetero, excluding Kimboza data).

The 10 most commonly mistnetted species (standard mistnetting; totally 847 individuals) were: Nectarinia loveridgei (243 specimens), Andropadus masukuensis (65), Pogonocichla stellata (57), Apalis thoracica (45), Cryptospiza reichenovii (42), Modulatrix stictigula (39), Phyllastrephus placidus (33), Alethe fuelleborni (32), Andropadus neumanni (29) and Trochocercus albonotatus (28).

The mistnet captures are the only good indicator of the abundance of some very elusive understorey species. Cryptospiza reichenovii is one of these. It was mistnetted often at most stations but was rarely recorded by observational methods or tape recordings. Also for Alethe fuelleborni the abundance is much underestimated by observational methods and tape recordings. This shy ground bird sings very infrequently and rarely gives warning calls. Mistnetting was furthermore a good method to describe the relative abundance of other more easily recordable understorey species, such as robins and greenbuls. A noticeable exception is the common species Bradypterus mariae which moves near the ground through the herbaceous vegetation and seldom flies into the nets. This vocal species is the second most common in the plot data.

Species foraging in the mid and upper strata are not prone to fly into nets, unless the forest is very low and gnarled. The mistnet results underestimate the relative abundance of certain species moving between strata, e.g. Trochocercus albonotatus (most common above the height of mistnets), although it was caught at most stations. The abundance of this vocal small flycatcher is much better described with the plot data. Other examples of discrepancy between mistnet and plot data are the above mentioned of Bradypterus mariae and the abundance figures for Andropadus masukuensis and A. neumanni at Kimhandu-3 (1940 m). These latter two species appear to be almost equally common judging from the mistnetting data whereas A. neumanni was recorded more than eight times as often as masukuensis during the plot assessments. The explanation is probably that neumanni is more vocal and more of an upper-strata species than masukuensis. For Nectarinia loveridgei, the relative abundance may be overestimated by mistnet data (compared to other species at the same stations) because it is very mobile and therefore very prone to fly into the nets. Mistnetting is ineffective where the forest is dry with the main insect activity being in the canopies as it appeared to be the case in Kimboza Forest during this survey (most species appeared to feed above 5 m where the insect activity was higher than in the dry forest floor).

The catch rate varied quite much between our stations, from 41.2 individuals per 2500 netmeterhours at Tegetero-3 (1710 m) to 92.5 individuals per 2500 netmeterhours at Kimhandu-1 (1520 m) (Table 5.8). It is interesting that the total number of species mistnetted was 15 at both these stations and that Kimhandu-1 is one of the stations with lowest average numbers of species per plot. The catch rate depends on a number of factors: general bird density, bird activity (dependent on e.g. the weather conditions, whether there are nestlings to feed or not, and how easy it is to fly fast through the vegetation), visibility of the nets (high visibility in case of sunny weather and open forest floor or no canopy cover, low visibility in case of dark forest and nearby vegetation) and dryness of the understorey (see above).

The number of species mistnetted generally stabilised some time during the second day (see Figure 5.4). For species-poor East African forests two days therefore seems to be a satisfactory minimum effort to put into the mistnetting with an effort of 2500 MNH = c. 100 m (if the nets are placed in different microhabitats and not in a long chain, same nets used both days). The number of individuals mistnetted per locality is quite small, however, and collected in a very small area. The data therefore gives only a very rough estimate and will not suffice for any robust statistical analysis between stations. J. Rabøl has demonstrated, using Monte Carlo simulations, that small samples give a false impression of a fairly constant species richness: samples of 50 birds drawn from three “communities”, each with 1000 individuals and a “normal” species abundance curve but 40, 66 and 99 species (S), respectively gave: four simulation with S=40: 18, 21, 22 and 22 species; with S=66: 22, 25, 27 and 31 species; with S=99: 25, 25, 30 and 31 species (Fjeldså and Rabøl in press). However, mistnetting remains useful as the only kind of data that can be collected by students who are not fully familiar with identifying birds by jizz and calls.

Mistnet data are presented in two different ways in this chapter. In Figure 5.1 the abundance of the individual species are presented as per cent of the total capture at the single station. The per cent abundance of a given species cannot be compared between stations because the figure depends on the abundance of other species, e.g. of very common ones like N. loveridgei (example: catching three Modulatrix stictigula per MNH at two stations where there are 97 and 197 other birds will yield relative abundances of 3 and 1.5 % respectively). The figures given as "individuals mistnetted per 2500 MNH" in Table 5.8 should be easy to compare with coming surveys if the mistnetting is standardised, but the figures admittedly depend on the general catch rate at the station (which depends on a number of factors, see above). For assessing the species diversity per station it would probably be best to ensure that at least 100 individuals were mistnetted per station and then give the species diversity as number of species per first mistnetted 100 individuals. This number would be easy to compare with also for other studies.

During this survey almost every bird was measured. This restricted the number of nets we could handle. When the primary purpose is to obtain comparative community data, intensions to measure all birds will represent a bottleneck. We hope that the attitude towards ringing without measuring all birds will change in Tanzania. Another solution would be to involve more people in the mistnetting process but this is not always possible. Logistics was admittedly also sometimes the limiting factor in the sense that we put up as many nets as time allowed us to at many of the stations.

One-hectare plots. With this method we recorded 55 forest species (plus Schoutedenapus myioptilus which we regard an aerial feeding forest species) and four non-forest species (Chrysococcyx klaas, Bradypterus cinnamomeus, Laniarius aethiopicus and Dryoscopus cubla) (Table 5.9).

The 10 most frequently recorded species on the totally 255 plots are Nectarinia loveridgei (180 plots), Bradypterus mariae (131), Andropadus neumanni (122), Trochocercus albonotatus (109), Modulatrix stictigula (97), Pogoniulus leucomystax (87), Apalis thoracica (87), Pogonocichla stellata (84), Andropadus masukuensis (65) and Turdus olivaceus (62).

For many species the plot data demonstrate trends which were also noted during general field observations. The method proved especially useful for describing the abundance of common birds foraging above the height of mistnets (e.g. Tauraco livingstonii, Ceratogymna brevis, Columba arquatrix, Apaloderma vittatum, Pogoniulus leucomystax and Apalis chapini) and of vocal understorey species (e.g. Modulatrix stictigula and Bradypterus mariae, both of which sing and give warning calls frequently). The plot data may to some extent underestimate the abundance of a very abundant species like N. loveridgei since we counted only constancy and not abundance per plot (for calculation of percentages in Table 5.1, five N. loveridgei on a plot counts the same as one individual on the plot). Of the single methods used on this survey, the one-hectare method is the most useful for rapid estimates of species richness and community composition.

The total number of species recorded per station depends not only on altitude and habitat complexity but also on the number of plots assessed. Due to time constraints (combined with often low bird activity), it was not possible to assess 25 plots at each station. The curves in Figure 5.4 show that the species accumulation has levelled out at stations where 22 or more plots were assessed (Kimhandu-2, -3 and -6; Tegetero-1, Tchenzema) and also at some stations where only around 15 plots were assessed (e.g. Kimhandu-4 and Lanzi-2). The best basis for a comparison is, however, still between stations with an equal effort and we advice future studies to allocate time for assessing 25 plots. On other stations it does not seem to have levelled out (e.g. at Kimhandu-1, Tegetero-2 and -3). Although the species accumulation has ceased, it remains to be tested to which extent the relative abundances of the species have stabilized after 25 plots and how the plot data correspond to absolute abundance measurements collected by territory mapping. Although 25 plots may provide a general characterisation of the community (the aim of making them), it is certainly not enough for recording the full assortment of low density species (like certain Threatened species ) in the area. In very species-rich tropical forests like certain lowland forests in South America, the species accumulation would probably level out more slowly than in the Ulugurus and it would be necessary to assess much more than 25 plots.

The number of species recorded per plot varied much, also within the same stations (see Figure 5.4), e.g. at Tegetero-1 (1345 m) and Kimhandu-3 (1940 m) where the lowest number of species recorded at a single plot is four, the highest 16. Probably the numbers recorded at the plots have not declined from the first to the last plot assessed in the morning.

Tape recordings at dawn. With this method we recorded 42 forest species (plus Schoutedenapus myoptilus which we regard an aerial feeding forest species) and one non-forest species (Bradypterus cinnamomeus).

Some of the species most often recorded at the stations are Tauraco livingstonii, Modulatrix stictigula, Pogonocichla stellata and Turdus olivaceus. Species with loud voices (e.g. Tauraco livingstonii) are often overrepresented on the recordings compared to species with less far-carrying voices, but if comparing only between stations and not with the results of other methods, this is not a problem.

The data generally support many of the trends that were obvious from mistnetting and plots. Furthermore it is the only of the standardised methods we used which is suitable for indicating the presence/non presence of species such as Accipiter tachiro, which generally vocalises only at dawn. The data sets are, however, very small and generally add little new information to the data obtained by mistnetting and one-hectare plots at the single stations. The tape recordings support the low number of species recorded at Kimhandu-5 (2520 m) in relation to other Kimhandu-stations but does not reflect the high number of species recorded by observational methods at the two lower Tegetero stations. The only example of a species who's presence at a station was revealed by tape recordings only is the presence of Camaroptera brachyura at Tegetero-1 (1345 m). Using 45 minutes of recordings per day might have improved the results but this was not always possible due to technical problems.

The method demanded little work in the field but the data analysis was quite time consuming. We cannot recommend this method for obtaining standardised data sets on future surveys in the Eastern Arc. However, if the fieldworkers are inexperienced with the vocalisations when arriving to an area, tape recordings can certainly be of much value for later identification of species. In extremely species rich tropical forests like those of the Brazilian lowlands, tape recording is an indispensable part of a survey since, at present, only around two to three persons can identify most bird voices while in the field (J. Fjeldså pers. comm.).

General field observations. 60 forest species (plus Schoutedenapus myoptilus which we regard an aerial-feeding forest bird) were recorded at the three main localities. A further 11 non-forest species were recorded inside forest in these areas (Appendix 5.6). Species recorded only in the ecotone between forest and surrounding cultivation are not included in the figures above. We also recorded a number of overflying species (swallows, swifts, raptors etc., see Appendix 5.6).

Several of the Threatened species known from the Ulugurus occur at so low densities here and in other parts of the Eastern Arc Mountains that they are not normally detected by methods such as mistnetting and timed plot assessments. To determine the status of such species much time must be allocated to undertake general field observations, visiting also remote and different parts of the forest habitat and observing many mixed feeding parties if such are present. For low density species such as Malaconotus alius it is essential to search wide areas. To record a low numbered species like Apalis chariessa (and probably Ploceus nicolli and some of the rare sunbirds) a large number of feeding parties must be studied.

During our survey we were under severe time pressure. Much time was used on mistnetting and one-hectare plot assessments. Also logistics (finding new camp sites, moving the large team, establishing campsites, putting up mistnets etc.) were very time consuming, although finding suitable campsites also means that considerable distances were walked with some bird watching being undertaken on the route (generally stops only when we heard or saw something interesting). Our survey of the status of Threatened species other than very audible and relatively common species like Scepomycter winifredae can only be considered thorough in the relatively small areas we covered though we attempted to survey as much as time allowed us to. We would have liked to have more time to search for M. alius and other Threatened species in the Uluguru North FR, e.g. by examining more mixed feeding parties. The value of searching thoroughly for Threatened species to obtain a good impression of their abundance cannot be overemphasized since Red Data Book species are a key factor for the determination of the conservation value of a forested area. The time constraint furthermore meant that at Lanzi and Tegetero we did not survey the entire gradients. See also notes in Section 5.4.4 on the efforts of general field observations and in Appendix 2 on the background of the survey.

Conclusion. All methods are biased to an unknown degree, and robust hypothesis testing therefore requires more exact, time-consuming methods (spot-/territory-mapping). However, the circumstances require rapid surveys of eastern African forests (key areas have to be identified quickly to initiate conservation programmes in time), and in this situation standardised data give a better basis for evaluating variations in community structure than qualitative data. Therefore a combination of mistnetting, plots (and perhaps tape recording) as described here seems to be a suitable approach. Much time must be allocated to general field observations to determine abundance and habitat preferences of Threatened, Near-threatened and other restricted-range species.

Appendix 5.13. Further recommendations for future ornithological fieldwork.

Mountain forests. Some obvious "holes" in the range of habitats which we succeeded in describing during this survey are the lowest parts of the Uluguru North F.R. (below 1300 m), the highest parts of the Uluguru North F.R. (above 1710 m) and small forest lobes like Kasanga F.R. (at c. 1000 m) in the southern section.

The large Lukwangule Plateau has been visited only briefly and there may be taxa in this interesting area, which we are not aware of, such as montane cisticolas.

The taxonomic status of the very distinct endemic subspecies of Uluguru Mountain Greenbul Andropadus neumanni needs further clarification. The southern satellite population of Cryptospiza salvadorii in the Ulugurus has not yet been assigned to subspecies. It is likely that it forms a separate, as yet undescribed, subspecies, but it is necessary to collect additional specimens to determine whether this is the case (see new data in e.g. Roy, M.S., da Silva, J.C., Arctander, P., García-Moreno, J. & Fjeldså, J. 1997. The role of montane regions in the speciation of South American and African birds. In D.P. Mindell (ed.) Avian Molecular Evolution and Systematics: 325-343. Acad. Press).

Lowland forests. It is important to take seasonal vertical migration into consideration when planning conservation initiatives in the Eastern Arc. The importance of lowland forests as dry season refuges for maintaining the populations of montane breeding species must not be forgotten, though the extent of the migration is still unknown. If these refuges are clearfelled or partly destroyed, it may affect also part of the submontane and montane avifaunas. Loss of forest on lower slopes and foothills may already have had an effect on the populations of species carrying out vertical migration in the Ulugurus.

Kimboza and Ruvu F.R.s are easily accessible from Dar es Salaam and would be a fine study site for Tanzanian students of biology to carry out standardised work on seasonal fluctuations in the habitat distribution of birds (for a discussion of these, see Appendix 5.8). Such a study could reveal whether the lowland forests are an important dry season refuge for populations of montane species. Additional small patches of forest in the lowland and some rich woodland areas should be visited at least briefly to check whether these types of habitat are of importance for montane species (the Andersen records of Pogonocichla stellata from the Ngerengere River indicate that this thrush may use a wide range of off-season habitats and it seems unlikely that the whole Uluguru population of P. stellata, if migratory, should depend solely on the relatively small areas of lowland forest. In addition, Terpsiphone viridis may move to woodlands in the cold season). Such visits to small lowland forest patches and rich woodlands (both habitat types are very threatened in the Ulugurus) would also determine whether these habitat types are of importance for the Near-threatened species Anthreptes neglectus and Circaetus fasciolatus.

---

[1] Nyange and Nyingwe are, according to Stuart and Jensen (1985), other names for Nyingwe (which is south of Lanzi,eastern slopes of the Uluguru South F.R., see Figure 1.4). However,Loveridge (1960) gives altitude (in feet) for some of the specimens collected in Nyange and Nyingwa, and all specimens listed from Nyingwa in his paper are from 2300 m above sea level, whereas all specimens from Nyange are from 750 m above sea level, indicating that the names must represent two different localities although probably in the same area.

[2] Stuart and Jensen (1985) suggested that the specimen described by Friedmann and Loveridge could be a misidentified Red-faced Crimson-wing C. reichenovii, possibly an immature bird, since the Ulugurus are well outside the known range of C. salvadorii (the nearest known locality being Kilimanjaro 400 km to the north) and since no other ornithologists had recorded the species in the Ulugurus. However, they must have overlooked Friedmann and Loveridge's citation from a paper by Shelley: "... in its most southern range the species has been met with by Dr. Stuhlmann at Uluguru." Furthermore Friedmann and Loveridge actually remarked that "This single specimen is obviously Cryptospiza salvadorii and not C. reichenowi". Britton (1980) did not mention C. salvadorii from the Ulugurus, neither did Hall and Moreau (1970). Our records confirm that Friedmann and Loveridge were right. Friedmann and Loveridge (1937) remarked that their specimen of C. salvadorii did not fit any known races. They designated the bird as "subsp. nov.?" and left the absolute determination until more specimens were available. Britton (1980) described the species as being a common resident (our remark: in its range to the north) of the undergrowth of highland forest and bamboo at 1500-3000 m, often in edge and secondary growth areas. It is necessary with further fieldwork (and collection of a few specimens) to determine its taxonomical status and habitat preference in the Ulugurus.

[3] Records of species seen in Morogoro Town or only below 900 m are available from the authors.

[4] A number of montane breeding Eastern Arc forest bird species are suspected to move to lower altitudes in the cold (dry) season. In the Usambaras the downward movement generally takes place from March to April, and the birds return in September-October to their breeding areas at higher altitudes (Stuart et al. 1993). The pattern of movement is probably rather much the same in other Eastern Arc ranges. Among some species, such as Pogonocichla stellata, probably the bulk of the population migrates to lower elevations (Stuart et al. 1993 and below) but other species migrate only on a small scale. In general the foothill forests and the lower parts of the submontane forests are believed to be of high importance for the migrants in the off-breeding season. Some individuals of very migratory species like P. stellata and Cercococcyx montanus (the migratory pattern of the latter probably being quite complex, but see below) do, however, sometimes migrate as far away as to the coastal forests in the dry season (Baker 1984; Britton et al. 1980).

During fieldwork in the Udzungwa Mountains 1991-92 (Ndundulu Mountains, forested from 1350 to 2400 m), very clear seasonal differences were observed for four species: Pogonocichla stellata, Terpsiphone viridis, Cercococcyx montanus and Smithornis capensis (altitudinal migration in the latter species was actually not mentioned by Keith et al. 1992). All four occured at very low densities when the field stations (1350, 1540, 1710 and 2140 m altitude) were visited in July-September (part of the dry season) but they were abundant when the fieldwork programme was repeated in November-January (season of short rains) (LAH and JOS). Also other species were most abundant in the latter field period but with a less pronounced difference, including Aplopelia larvata, Zoothera gurneyi, Andropadus olivaceiceps, Apaloderma vittatum and Muscicapa adusta. A similar pattern may be found in the Ulugurus. It should be stressed, however, that for a given species the pattern and extent of migration may well differ between mountain ranges (see Dowsett-Lemaire 1989 for a discussion of this, based mainly on Malawian montane forests but including also results from other Afromontane forests). The pattern may furthermore change from year to year.

Stuart et al. (1993) briefly summarized the knowledge on altitudinal migration in the Eastern Arc. The following list is compiled from their paper supplemented with data from the Cambridge Tanzania Rainforest Project (1994). For some of the species the assumption is based on very few observations from lowland forest or from low altitude submontane forest in the off-breeding season.

List of strictly montane breeding species known or believed to undertake seasonal vertical migration in the Eastern Arc:

(Ul = Ulugurus; Ud = Udzungwas; Us = Usambaras).

Columba arquatrix (Ud), Aplopelia larvata (Ul, Ud), Cercococcyx montanus (Ul, Ud, Us), Apaloderma vittatum (Ul, Us), Dendropicos griseocephalus (Us), Alcippe abyssinica (Us), Andropadus masukuensis (Us), A. neumanni (Ul, Us), A. olivaceiceps (Ul, Us), Alethe fuelleborni (Ul, Us, Ud), Modulatrix stictigula (Ul), Sheppardia sharpei (Us), Pogonocichla stellata (Ul, Us, Ud), Zoothera gurneyi (Ud, Us), Trochocercus albonotatus (Ul, Us), Telophorus nigrifrons (Ul, Ud, Us), Onychognathus walleri (Us), Poeoptera kenricki (Ul, Us), Anthreptes rubritorques (Us) and Cryptospiza reichenovii (Us).

Some montane breeding species which also have breeding populations in the lowlands or in non-forest habitats (wherefore they are not included in the list above) are also believed to migrate altitudinally to various extents. They include Terpsiphone viridis (regarded an “intra-African migrant” at least in parts of Africa), Smithornis capensis (see comments above under the Udzungwa Mts.), Black Cuckoo Shrike Campephaga flava (regarded an “intra-African migrant” at least in parts of Africa), Coracina caesia and Muscicapa adusta (e.g. Stuart 1983 and above) .

A three weeks general survey in the large lowland Matundu Forest (at the base of the Udzungwas, c. 400 m) in September 1994 revealed that Pogonocichla stellata was abundant whereas the rest of the assemblage of species suspected to migrate altitudinally were recorded only in low numbers (T. Lehmberg pers. comm.). Of the above-mentioned species the following were recorded: Cercococcyx montanus (some singing, not many), Coracina caesia (one seen), Campephaga flava (fairly common), Smithornis capensis (some singing), Zoothera gurneyi (one mistnetted), Alethe fuelleborni (possibly heard at 2-3 occasions), Telophorus nigrifrons (some seen). Matundu was dry at the time of the visit. Dryness may be a limiting factor on the amount of food available (in Matundu as well as in Kimhoza).

The seasonal altitudinal migration has never been studied in detail in the Eastern Arc, and the knowledge about this phenomenon is still very incomplete. It may turn out that some of the species, who's populations in the Eastern Arc are regarded as strictly montane breeding, have a scattered occurrence in the lowlands throughout the year as is the case in other parts of Africa for some of them. Some of the occurrences in lowlands may also best be explained by exploratory movements - in Malawi, e.g. Apaloderma vittatum, Telophorus nigrifrons, Aplopelia larvata and Coracina caesia have been shown to undertake exploratory inter-forest movements, the first two species up to 100 km, the last two up to 80 km, through inhospitable terrain (Dowsett-Lemaire 1989). Some species of the Malawian forests undertake intra-African migration, which cannot be termed altitudinal migration (Dowsett-Lemaire 1989).

[5] From Table 5.5 it is clear that species have been overlooked at some of the stations. Nonetheless it is possible to compare the stations considering the total number of species recorded, with some care.

[6] Dryoscopus cubla is generally a species of forest edges (for which reason it was maybe best to treat it as a forest species).

[7] Laniarius aethiopicus occupies a range of habitats.

[8] This species is best recorded with mistnetting but no mistnetting has been carried out at Tchenzema.