Phylogeography of Ocean Striders
Σ Copyright, N. M. Andersen & Zoological Museum, University of
species of sea skaters, Halobates
germanus, H. micans, H. sericeus, H. sobrinus and H.
splendens have successfully colonized the open ocean, where adults and
juveniles spend their entire life on the sea surface, always at some distance
from land. They feed on other animals belonging to the pleustonic community and
are themselves preyed upon by seabirds and pelagic fish. Eggs are deposited on
various floating objects, sometimes even on live heteropods.
distributions of the five oceanic Halobates species are now well
established by Cheng (1989). H.
micans is widely distributed in the Atlantic, Indian, and Pacific Ocean. H.
sobrinus and H. splendens are confined to the eastern Pacific Ocean. H.
germanus occurs in the Red Sea, Indian Ocean, and western Pacific Ocean.
Finally, H. sericeus is found throughout the Pacific Ocean except for a
and south of the equator which is occupied by H. micans. Sea skaters only
occur in tropical and subtropical waters where the winter temperature does not
fall much below 20o C. Their abundance on any given water mass was
found to be correlated with sea-surface temperatures with an optimum temperature
range for the four eastern Pacific species (H. micans, sericeus, sobrinus,
and splendens) of 24-28o C.
Andersen et al. (2000)
investigated the genetic variation within the five oceanic Halobates
species and estimated the gene flow between
populations inhabiting more than one ocean (H. micans and H. germanus)
or different parts of the same ocean (H. micans, H. sobrinus, H. splendens,
and H. sericeus).
Nucleotide variation and divergence
The table (above) summarizes the results of
the studies of the genetic variation within of the oceanic species of Halobates.
In H. micans sampled from the Atlantic, Pacific and Indian oceans (n =
29), a total of 85 nucleotide sites vary over a 780 bp region of COI mtDNA. The
intraspecific variation of H. micans (05.00%) is much larger than that
observed among specimens of other species of Halobates, e.g., 00.77%
in 13 specimens of H. nereis (Palau Islands) and 01.54% in 16
specimens of H. germanus (Indian and West Pacific oceans), and is more
than half of the variation found when pairs of closely related species of Halobates
are compared, e.g., H. micans and H. splendens (4.496.82%), H.
flaviventris and H. hawaiiensis (6.41%), and H. germanus and H.
(p) nucleotide divergences are higher for pair-wise comparisons among Halobates
micans specimens from the Indian Ocean (Mean = 1.41%, SE = 0.37%, N = 45)
than among specimens from the
Atlantic Ocean (Mean = 0.92%, SE = 0.27%, N = 36) or from the Pacific Ocean
(Mean = 0.81, SE = 0.32%, N = 45). Much higher nucleotide divergences was
obtained from pair-wise comparisons between Halobates micans specimens
from different oceans. Specimens from the Atlantic and Pacific Ocean differ by
2.573.98% (Mean = 3.44%, SE = 0.27%, N = 90), specimens from the Atlantic and
Indian Ocean differ by 3.335.00% (Mean = 4.16%, SE = 0.37%, N = 90), and
specimens from the Pacific and Indian Ocean differ by 3.085.00% (Mean =
4.19%, SE = 0.43%, N = 100).
The possible association between nucleotide
divergence and geographical distance was investigated for pairs of Halobates
micans specimens from the three major oceans. Distances range from about 300
to about 6,300 km in the Atlantic Ocean (see map below), from less than 100 km
to more than 15,000 km in the Pacific Ocean, and from less than 100 km to about
1,800 km in the Indian Ocean. There is only weak association between nucleotide
divergence and geographical distance. In the Pacific Ocean, specimens with
identical or very similar haplotypes (00.13%) were sampled from localities
less than 500 km apart as well as in localities more than 15,000 km apart.
nucleotide variation in Halobates splendens (n = 6) and H. sobrinus
(n = 8), both from the eastern Pacific Ocean, are 1.4% and 0.26%, respectively.
The range of nucleotide divergence between pairs of specimens of these two
species are 01.15% (Mean = 0.73%, SE = 0.44%, N = 15) and 00.26% (Mean =
0.09%, SE = 0.08%, N = 28), respectively. Geographical distances between
specimens range between 100 and 3,000 km, without any marked association with
In Halobates germanus sampled from the southern Pacific (n = 3) and western Indian Ocean (N = 13) a total of 19 nucleotide sites vary over the 780 bp region of COI mtDNA. Nucleotide divergence among specimen-pairs of H. germanus are generally smaller than in H. micans except for a 0.511.67% divergence between specimens from the Pacific and Indian Ocean. Within the western Indian Ocean, nucleotide divergences between pairs of specimens are 00.90% (Mean = 0.31%, SE = 0.22%, N = 78) and geographical distances vary from less than 100 km to about 1,400 km without significant correlation between nucleotide divergences and geographical distance.
Halobates sericeus from the northern and southern Pacific Ocean (n
= 8), a total of 11 nucleotide sites (1.4%) show variation. The range of
nucleotide divergence between pairs of specimens is
01.41% (Mean = 0.80%, SE = 0.68%, N = 28). The geographical distance
between pairs of specimens range between 100 to about 8,500 km. It is noteworthy
that specimens with the same haplotype (se1) were sampled from localities
5308,300 km apart.
Halobates micans from the Indian, Pacific and Atlantic oceans. Neighbor-joining
within Halobates micans
data set composed of 27 haplotypes of Halobates micans was analyzed using
both neighbor-joining and maximum parsimony methods, with H. splendens
and H. sobrinus as outgroups. Both methods keep together all H. micans
haplotypes and furthermore distinguish groups of haplotypes from the Atlantic,
Pacific, and Indian Ocean, each supported by relatively high bootstrap values
(98, 83, and 99%, respectively; see above) as well as branch support (10, 3, and
4, respectively). Haplotypes from each of the major oceans occupied by H.
micans are significantly different, having sequences containing 57 unique
base substitutions, including two nonsynonymous substitutions which, translated
to proteins, imply one amino acid difference only found in haplotypes from
the Indian Ocean and one found in specimens from the Atlantic and Pacific
Oceans. In terms of pair-wise similarity between COI-sequences, haplotypes from
the Atlantic Ocean are more similar to those from the Pacific Ocean than both
are to haplotypes from the Indian Ocean.
that such differences in COI-sequences reflect present genetic isolation between
populations and, indirectly, the
time passed since the last separation of populations inhabiting major ocean
basins, it is hypothesized that populations of Halobates micans now
inhabiting the Indian Ocean (or at least its western parts) became isolated from
other populations of this species before the separation of populations
inhabiting the Pacific and Atlantic Oceans. The observed genetic differences
between these populations are, however, not correlated with morphological
differences that justify a subspecific or specific differentiation of H.
micans (Andersen et al., unpublished observations).
a nucleotide substitution rate of 1.1 to 1.2% per Myr (as commonly used for
arthropod mtDNA) to the observed nucleotide divergences for Halobates micans,
the Atlantic populations may have been separated from the Pacific populations
1.11.8 Myr ago and from the Indian Ocean populations 1.4-2.2 Myr ago, or about the same time
when the Pacfic and Indian Ocean populations became separated. On the other
hand, if the minimum age of separation between populations inhabiting the
Atlantic and Pacific oceans is 3 Myr (as dated by the emergence of the Isthmus
of Panama), the estimated substitution rate in Halobates is 0.40.7%
per Myr or only about half the value estimated above. The most probable
explanation for this deviation is that Halobates diverge at a slower rate
than the sample of arthropods used to calibrate the molecular clock for
arthropod mtDNA. Another, although less probable explanation is that the final
closure of the Isthmus of Panama is of later date.
relatively large nucleotide differences between Halobates micans
haplotypes found in the Indian and Pacific Oceans (3.15.0%) suggests that the
Indo-Malayan Archipelago constitutes a barrier for continuous gene flow between
populations in these two oceans.
Based upon a theoretical investigation of the effects of oceanic diffusion on the distribution of oceanic Halobates, Ikawa et al. (1998) showed that an estimated radius of a "patch" of Halobates could be expanded by oceanic diffusion alone from an initial point of origin to 1,250 km in 60 days. This distance is only 1/12 of the maximum distributional range of H. micans in the Pacific Ocean. Since eggs of oceanic Halobates may take more than one month to hatch, nymphal development may take up to 70 days, and adults can live for at least two months, the descendants of a pair of individuals may be carried across the Pacific Ocean within the time required for several life cycles. Following these calculations, Ikawa et al. (1998) hypothesized that gene flow could occur by oceanic diffusion alone over the whole distributional range of a Halobates species. The results of the present investigation of nucleotide variation in oceanic Halobates, in particular those for H. micans, do not contradict this hypothesis.
Halobates splendens, the sister species of H. micans, is
restricted to the southeastern part of the Pacific Ocean, H. micans
probably originated somewhere in the Indo-Pacific and subsequently spread to the
Atlantic Ocean. Several possible routes of dispersal have existed in the past
(Jaczewski, 1972; Andersen, 1999). In
the Early and Middle Eocene, the Mediterranean was connected with the
Indo-Pacific Ocean by the Tethys Sea, located in present day Turkey, Iraq, and
The Persian Gulf. Although the finding of a fossil Halobates in 45 Myr
old, marine deposits in northern Italy
(Andersen et al., 1994) indicates the presence of sea skaters in
the Mediterranean as early as Middle Eocene, the ocean skaters probably evolved
more recently than that.
have been other, more recent routes of dispersal to the Atlantic Ocean, e.g.,
from the Indian Ocean around southern Africa, or from the Pacific Ocean through
Mesoamerica, before the closure of the Isthmus of Panama, about 3 Myr ago. At
present, dispersal of oceanic Halobates around the southern end of Africa
is hampered by the presence of cold surface currents along the West coast of the
African continent. On the other hand, surface temperatures in these parts of the
Atlantic Ocean may have been higher in the past. However, the magnitude of
nucleotide divergences between H. micans haplotypes from the Indian and
Atlantic Ocean definitely speaks against a recent gene exchange between
populations by dispersal around the southern end of Africa.
Andersen NM. 1999. The evolution of marine insects: phylogenetic, ecological and geographical aspects of species diversity in marine water striders. Ecography 22: 98-111.
Andersen NM, Cheng L, Damgaard J, Sperling FAH. 2000. Mitochondrial DNA sequence variation and phylogeography of oceanic insects (Hemiptera: Gerridae: Halobates spp.). Marine Biology 136: 421-430.
Andersen NM, Farma A, Minelli A, Piccoli G. 1994. A fossil Halobates from the Mediterranean and the origin of sea skaters (Hemiptera, Gerridae). Zool. J. Linn. Soc. 112: 479-489.
L. 1989. Factors limiting the distribution of Halobates species. In
Ryland, J. S. & Tyler, P. A. (eds.), Reproduction, genetics, and
distributions of marine organisms.
Olsen & Olsen, pp. 357-362.
Ikawa T, Okubo A, Okabe H, Cheng L. 1998. Oceanic diffusion and the pelagic insects Halobates spp. (Gerridae: Hemiptera). Marine Biology 131: 195-201.
Jaczewski T. 1972. Geographical distribution of oceanic Heteroptera and the continental drift. Bull. Acad. Sci. pol., Ser. Biol. 201: 415-417.
|Back to:||Top of page||Main page|
Last update: 04 september 2002