BREVIA
These changes in inferred trophic ecology are
significantly correlated with evolutionary changes
in armor phenotype through time (3) (Fig. 1E).
DF scores are correlated with dorsal [nonparametric Spearman rank correlation (rs) = 0.23, P =
0.03, n = 89 fish] and pelvic armor (rs = 0.21, P =
0.05), feature density with dorsal armor (rs = 0.24,
P = 0.02). Interestingly, the shift to a more benthic
Mark A. Purnell,1* Michael A. Bell,2 David C. Baines,1 Paul J. B. Hart,3 Matthew P. Travis2
ecology within sample 19.8 (Fig. 1, D and E)
odels, experiments, and field studies high-resolution record of evolutionary change precedes the increase in mean armor scores in 19.6
provide evidence of the ecological within a lineage spanning tens of thousands of (a time lag of circa 100 years). This evidence of an
ecological shift preceding phenotypic change
controls on evolution, but extrapolating years (3).
We investigated the relationship between troph- suggests that this part of the sequence may record
results over longer time scales is a perennial
problem in evolutionary biology. Trophic ecology ic resource use and evolutionary change through rapid evolution driven by shifts in trophic ecology
and competition for food, for example, are thought quantitative analysis of dental microwear (4). and adaptation to benthic niches. If this hypothesis
to drive speciation through niche differentiation, Laboratory feeding experiments and analyses of is correct, however, the low number of specimens
character displacement, and phenotypic wild stickleback populations show that microwear displaying intermediate phenotypes is puzzling,
divergence (1). Yet direct evidence that feeding exhibits a progressive shift from planktivores to and the scenario of replacement of one lineage by
controls evolution over extended time scales, benthic feeders (Fig. 1, A and B) (5). Discriminant another (3) cannot be ruled out. The gradual shift
to less benthic ecology over the
next 17,000 years supports the
2.9
Mud (94%B; 17.8) C
ssample 19.6 50
D E 20
A
3.8
50 B
Corcoran (76%B; 20.4)
19.7
interpretation that a return to lowKashwitna (51%B; 20.9)
19.8 40
armor phenotypes reflects direcLong (47%B; 21.9)
6.9
6.9
40
3.8 30
tional natural selection (3).
15
2.9
21.5 20
Our analysis shows that dental
30
10
microwear analysis can provide
10
direct evidence for changes in
20
0
1 2 3 4
trophic niche and resource exploiBenthic-Planktivore
Benthic-planktivore
10
trend
tation in fossil fishes. That changes
19.6
5
19.7
in feeding can be detected in19.8
21.5
0
dependently of morphological
2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
1
2
3
20
25
30
change highlights the potential of
Dorsal spine / pelvic score Mean feature density
Feature length (µm)
this approach to provide important
Fig. 1. Microwear in stickleback teeth and correlated evolutionary change. (A) Scanning electron micrographs
showing tooth microwear in fossil (top) and extant benthic feeding (bottom) stickleback. For details, see fig. S1. insights into trophic ecology dur(B) Microwear in wild-caught and lab-raised stickleback. Open ellipses indicate lab distributions (blue, benthic ing adaptive radiations of fishes
treatments; red, planktivore; solid, dashed, and dotted lines are course, medium, and no sand substrate, and other evolutionary events.
respectively). In wild fish, microwear tracks trophic ecology as indicated by % benthic stomach contents and mean
References and Notes
gill raker count (Mud Lake, most benthic; Long Lake, least benthic). (C) Fossil stickleback microwear; inset (D)
1. D. Schluter, The Ecology of Adaptive
shows sample 19.8 divided into earlier (1746 to 1753 years) and later (1757 to 1771 years) subsamples with shift
Radiations (Oxford Univ. Press,
toward more benthic trophic ecology in later interval. Ky, thousand years. (E) Trophic niche and morphology in
Oxford, 2000).
fossil stickleback through time [○ dorsal armor; + pelvic armor; ◊ mean feature density; × DF scores (minimum of
2. D. Schluter, J. D. McPhail, Am. Nat.
140, 85 (1992).
0.38 and maximum of 2.51)]. Colored horizontal bars show niche scores reflecting the position of the samples in
3. M. A. Bell, M. P. Travis, D. M. Blouw,
the benthic-planktivore microwear spectrum (C). Time scale follows (3).
Correlated Evolution and Dietary
Change in Fossil Stickleback
available only from the fossil record, is difficult
to obtain because it is rarely possible to directly
analyze dietary change in long-dead animals.
Functional changes must be inferred from changes
in morphology, and attempts to determine whether
morphological changes were caused by shifts in
feeding can become circular.
Here, we report an investigation of trophic
resource use in a fossil sequence preserving an
evolving lineage of threespine stickleback
(Gasterosteus). We focus on stickleback for
two reasons. First, perhaps the best-known work
on speciation in fishes concerns stickleback in
postglacial coastal lakes in Canada, where
planktivores and benthic feeders coexist as two
reproductively isolated and phenotypically distinct trophic forms. The differences between these
forms result from competition for food (1, 2).
Second, fossil stickleback from the Miocene
Truckee Formation (Nevada) provide a detailed,
analysis using feature length and density indicates
that scores for the first discriminant function (DF)
are a good predictor of trophic ecology. For wild
fish populations (n = 4), mean scores were significantly correlated with diet (r = 0.95, P = 0.05)
and gill raker number (r = –0.996, P = 0.004).
Analysis of fossil stickleback teeth revealed an
overall range and pattern of feature densities and
lengths similar to that of extant fish (Fig. 1C), suggesting that the fossil microwear records a similar
benthic-planktonic feeding spectrum. This was
supported by application of the DF derived from
wild fish to the fossils: DF scores vary significantly
between samples (F = 10.8, df of 7 and 87, P =
0.0001), and a Tukey-Kramer procedure revealed
significant pairwise differences. This procedure also
grouped some fossil samples with benthic-feeding
wild populations (samples 19.6, 19.7, 19.6, and
6.9), others with planktivore populations (21.5),
with some placed between (2.9 and 3.8).
www.sciencemag.org
SCIENCE
VOL 317
more benthic
Time (Ky)
earlier
Feature density (mm-2) /1000
later
more limnetic
M
Paleobiology 32, 562 (2006).
4. Materials and methods are available
as supporting material on Science Online.
5. M. A. Purnell, P. J. B. Hart, D. C. Baines, M. A. Bell,
J. Anim. Ecol. 75, 967 (2006).
6. Funded by Natural Environment Research Council (NERC)
grants NER/B/S/2000/00338 and NE/B000125/1 to M.A.P.
and P.J.B.H.; M.A.P. also supported by NERC Advanced
Fellowship NER/J/S/2002/00673. Fossil stickleback
collected and prepared under NSF grants EAR9870337
and DEB0322818 to M.A.B. and F. J. Rohlf.
Supporting Online Material
www.sciencemag.org/cgi/content/full/317/5846/1887/DC1
Materials and Methods
Fig. S1
3 July 2007; accepted 15 August 2007
10.1126/science.1147337
1
Department of Geology, University of Leicester, Leicester LE1
7RH, UK. 2Department of Ecology and Evolution, Stony Brook
University, Stony Brook, NY 11794–5245, USA. 3Department
of Biology, University of Leicester, Leicester LE1 7RH, UK.
*To whom correspondence should be addressed. E-mail:
[email protected]
28 SEPTEMBER 2007
1887
Supporting Online Material
Materials and Methods
For details of analysis of teeth from laboratory and wild-caught fish see ref (1). The fossil teeth were
collected by one of us (MPT) from a section through the Miocene Truckee Formation of Nevada (2). In
the laboratory at Stony Brook, a fine moistened paint brush was used to extract teeth from articulated
fish collected from seven intervals (see ref 2 for details of collecting and dating fish). Each interval
spans roughly 100 years except for samples 19.7 and 19.8. These are narrower in order to capture
changes during the interval of most rapid morphological change. In years from base of section samples
are: sample 21.5 = 88 - 207 yrs (estimated); sample 19.8 = 1,746 - 1,771 yrs; sample 19.7 = 1,776 1,828 yrs; sample 19.6 = 1,863 - 1,978 yrs; sample 6.9 = 14,456 - 14,635 yrs; sample 3.8 = 17,705 17,813 yrs; sample 2.9 = 18,595 - 18,701 yrs. After initial analysis revealed that samples below 19.8
exhibit more planktonic microwear and those above more benthic, this sample was further subdivided
into earlier (1,746 - 1,753 yrs) and later (1,757 – 1,771 yrs) subsamples to determine whether changes
in microwear occurred within it.
In sampling teeth, fish size was not an explicit criterion for selection, but the size of the
majority of fishes in the collections indicates they were adults, and the temporal trends in microwear
results cannot reflect ontogenetic diet change. That tooth size (as recorded by the area analysed for
microwear) does not vary between samples (ANOVA, F = 0.77, d.f = 6, 88, P = 0.60) provides further
support for this.
Teeth were sent to Leicester in micropalaeontological slides, and after cleaning in acetone and
coating with silver, scanning electron micrographs of an 80 x 100 µm area of the distal portion of the
labial surface of one dentary tooth from each fish in each sample were obtained using a Hitachi S-520
(Fig. S1). Microwear features were scored for the distal most 75µm and summary data calculated using
the custom software package Microware 4.02 (3, 4). This is a semi-automatic method, requiring
operator input to determine microwear feature boundaries. All microwear data were acquired by the
same operator (DCB); analysis of operator error indicates that replicate data sets acquired by DCB over
a period of weeks did not differ significantly from one another (1). We focussed on microwear feature
density and feature length, which for trophic analysis of stickleback are the most informative of the
data generated by the Microware software. For further details of our methods, see ref (1). Analysis of
tooth microwear and trophic interpretation was carried out blind: the investigators who scored and
interpreted microwear did not know the relative stratigraphic positions of the fossil fish horizons
sampled until the analysis and trophic interpretation were complete. Because pelvic armor phenotypes
are scored using an ordinal scale, correlations between microwear and armor were tested using nonparametric Spearman Rank Correlation (rs). Density ellipses in Fig. 1B-C, intended only as a guide to
the pattern of distribution of the data, are drawn to include 80% of the data for each sample, assuming a
bivariate normal distribution (sample 21.5 extreme outlier excluded).
It is unlikely that changes in sediment are influencing fossil microwear to any significant degree
because although experimental data indicate that microwear in treatments with course and medium
sand substrates differs from those with fine sand or no substrates (1) there is no evidence for grain size
changes of this magnitude in the Truckee Formation sequence (5). Difference between planktivore and
benthic microwear in the fossils is unlikely to be the result of increased variance resulting from time
averaging. Compared to samples exhibiting benthic microwear patterns, samples with planktivore
microwear have lower minimum and lower maximum values for feature density (i.e. less variance),
combined with greater range and higher values for feature length; this is not what would be expected
from time averaging (6). Furthermore, recent work has shown that the effects of time averaging on
phenotypic variance are less than has been thought, and that data are generally comparable to their
population-level equivalents (7). Some of our samples have very different durations but the variance of
the microwear data does not differ significantly. For example, samples 6.9 and 19.8 have similar
microwear patterns, both falling between the mid-range and the benthic end of the spectrum. Results of
ANOVA indicate that microwear data is not significantly different (for feature density and length,
respectively, F = 0.78, P = 0.38; F = 0.02, P = 0.88, d.f. 1, 37 in both cases), even though sample 6.9,
spans 179 years, and sample 19.8 spans only 25 years.
The data for stomach contents of wild fish populations shown in Fig. 1 indicate trophic niche, but the
mean of percentage food types found in each fish cannot capture some important aspects of the pattern
of variation. For example, although by any measure it is the most planktivorous of the populations
analysed, fish from Long Lake have a bimodal distribution of stomach contents (i.e. 30% of fish had
100% planktonic food items, 25% of fish had 100% benthic items). This is probably why the
distribution of microwear data differs from the laboratory planktivore treatments. Kashwitna Lake fish
were genuinely mixed feeders (20% of fish had 100% benthic items, the remainder spread across the 095% range). Fifty percent of Corcoran Lake fish had 100% benthic food items, and the remainder were
spread. 80% of Mud Lake fish had 100% benthic prey, and of the remainder, none had less than 35%.
Fish from Lynda Lake (1) exhibit no preferences regarding plankton/benthic food (no more than 11%
of fish in any % prey bin) and were not included in the present analysis.
Figure S1. Scanning electron micrographs of Miocene and recent stickleback teeth. A. Tooth from
fossil benthic sample 19.7 (image no. 04808). B. Tooth from wild benthic population from Mud Lake
(image no. 92109). C. Tooth from laboratory benthic treatment, course sand substrate (image no.
84403). D. Tooth from fossil planktivore sample 21.5 (image no. 00832). E. Tooth from wild
planktivore population from Long Lake (image no. 94309). F. Tooth from laboratory planktivore
treatment, medium sand substrate (image no. 81106). Wild fish were obtained from lakes in the
Matanuska-Susitina Valley, Alaska . Fossil fish were obtained from the Miocene Age Truckee
Formation, Nevada. Scale bar 50 µm.
Supplementary references
S1.
M. A. Purnell, P. J. B. Hart, D. C. Baines, M. A. Bell, J. Anim. Ecol. 75, 967 (2006).
S2.
M. A. Bell, M. P. Travis, D. M. Blouw, Paleobiology 32, 562 (2006).
S3.
P. S. Ungar. (Privately published, Fayetteville, Arizona, USA, 2001).
S4.
P. S. Ungar, Scanning 17, 57 (1995).
S5.
M. A. Bell, in Evolutionary Biology of the Threespine Stickleback M. A. Bell, S. A. Foster, Eds.
(Oxford University Press, Oxford, 1994) pp. 438-471.
S6.
M. A. Bell, M. S. Sadagursky, J. V. Baumgartner, Palaios 2, 455 (1987).
S7.
G. Hunt, Paleobiology 30, 487 (2004).