Supplementary material
Detailed description of experimental procedure and data collection
Females of one population were introduced into the aquarium for acclimatization one week prior to the introduction of males. Before introduction into the aquarium, male size (standard length to the nearest 0.1 mm) and male weight (to the nearest 0.1 g) were recorded. Males within a same-population pair were size matched (allowing a maximum deviation of 8% of the mean size). Furthermore, we tried to minimize the size differences between the male-pairs of the three P. nyererei populations and of P. igneopinnis (see below). Due to practical constraints on the number of males that could be used in the experiments no definite maximum size difference was defined between male-pairs (see below). We did not aim at maximizing or minimizing differences in components of coloration between males within a pair and hence chose to match males based only on size (i.e. irrespective of the extent of differences in coloration). Size and weight were recorded a second time after a replicate was terminated and the means of before and after were used for further analyses. An experimental replicate continued until a female had spawned. This lasted between 1 and 21 days (mean ± standard deviation = 7 days ± 6). The aquarium was inspected at least twice a day (mornings around 9 a.m. and afternoons around 4 p.m.) for spawnings. As soon as at least one female had spawned, all males were removed on the same day. One male at a time was removed and immediately photographed with fully spread fins in a Plexiglas cuvette. Afterwards males were anesthetized in a solution containing 0.5g/l MS222 and 0.25g/l baking soda and once no gill cover movement was detected they were injected with a 12-mm PIT-tag for individual identification (Biomark; Boise, ID, USA) and a piece of fin tissue was taken for DNA. Five days after the spawning event, all females that spawned were removed from the aquarium and a new set of males was introduced. The number of females within a replicate for all three population mate choice experiments ranged between 15 to 30. Whenever possible the number of females that spawned and that were removed after 5 days post-spawning were replaced with new, non-gravid females that were not previously used in the experiment. A unique combination of males was established for each replicate. The total number of males used over all mate choice experiments were N=32 for Python, N=17 for Senga and Makobe and N=14 for P. igneopinnis. The fry were gently removed from each of the females’ mouth and afterwards the females were anesthetized (see above) and a small piece of fin tissue for DNA was taken. Afterwards we recorded the weight and size of the females and then they were returned to the stock aquaria and not used again. The eggs were transferred to and raised in identical egg tumblers until day 14 post-spawning. Subsequently, the larvae were sacrificed (anesthetized (see above) and then euthanized in a water solution containing 1g/l MS22 and 0.5g/l baking soda) and subsequently preserved in 95% ethanol. Males and females used in the mate choice experiment PNM_1 were 2-4 years of age. Males and females from the other mate choice experiments (PNM_2, PNP and PNS) were 1-2 years of age. The average weight and size of all females and males in the experiments are given for each population in Table S3 and the differences in the weight between males in Table S4.
Figure S1
Representative photos depicting the variation in color observed in the experimental males used from the three P. nyererei population (Python, Senga, Makobe) and from P. igneopinnis.
P. nyererei males from Python are bright yellow between the bars on the lower flanks and orange to red above the lateral line. The dorsal, caudal and anal fins are orange to red. The anterior part of the dorsal fin often shows some sky blue iridescence. The ventral side of the flanks from the snout to the caudal fin can range from sky blue to light black. The light (and rarely dark) black bars along the flanks usually do not exceed the lateral line.
P. nyererei males from Senga are bright yellow between the bars on the lower flanks and show a range of colors from orange, red to magenta above the lateral line. The dorsal fin shows the same range of orange, red and magenta and the anterior part of the dorsal fin often has some sky blue iridescence, which can extend towards the caudal fin. The caudal and anal fins are orange to red. The ventral side of the flanks from the snout to the caudal fin are light to dark black (and rarely sky blue) and the black bars sometimes extend above the lateral line on the dorsum of the fish.
P. nyererei males from Makobe are bright yellow between the bars on the lower flanks and bright crimson red above the lateral line. The dorsal fin is bright crimson red and the caudal and anal fins are orange to red. Very rarely does the dorsal fin show some sky blue iridescence in its spinous part. The ventral side of the flanks from the snout to the caudal fin is dark black and the black bars often extend above the lateral line on the dorsum of the fish.
The males of P. igneopinnis are nearly entirely velvet black but have nyererei-like bright orange to red on the dorsal half of the snout. The dorsal fin is black with some blue iridescence in the spinous parts and the edge (lappets) and posterior part is bright orange. The caudal and anal fins are largely bright orange and the pelvic fin is entirely orange.
Table S1.
Backward model selection. In two cases the backward linear model selection had to be complemented by forward or "forced" backward model selection (see Methods and Results section for details). In the first column the specific test is reported (intrapopulation sexual selection or interpopulation assortative mating). In the second column the population of the females that were tested in the mate choice experiments is given together with the mode of model selection that was used. The model with the lowest AIC is the most likely model and marked in bold.
Test / Mate choice experiment / Model selection / Model / AICintrapopulation / PNM / backward selection / mate choice = red + orange + magenta + yellow + blue• + black / 32.5
intrapopulation / PNM / backward selection / mate choice = red• + orange + magenta + blue* + black / 30.5
intrapopulation / PNM / backward selection / mate choice = red* + orange + blue* + black / 28.8
intrapopulation / PNM / backward selection / mate choice = red* + blue* + black / 28.5
intrapopulation / PNP / backward selection / mate choice = red + orange + magenta + yellow + blue + black / 31.5
intrapopulation / PNP / backward selection / mate choice = red + orange + magenta + yellow + black / 29.5
intrapopulation / PNP / backward selection / mate choice = red + orange + yellow + black / 27.5
intrapopulation / PNP / forced backward selection / mate choice = red* + yellow** + black* / 42.2
intrapopulation / PNP / forced backward selection / mate choice = red + yellow* + orange• / 43.2
intrapopulation / PNP / forced backward selection / mate choice = red + orange + black / 63.6
intrapopulation / PNP / forced backward selection / mate choice = orange* + yellow* + black• / 50.7
intrapopulation / PNS / backward selection / mate choice = red + orange + magenta + yellow + blue + black / n.a.
intrapopulation / PNS / forward selection / mate choice = red / 27
intrapopulation / PNS / forward selection / mate choice = orange* / 16.8
intrapopulation / PNS / forward selection / mate choice = magenta• / 25.7
intrapopulation / PNS / forward selection / mate choice = yellow / 18.4
intrapopulation / PNS / forward selection / mate choice = blue / 26.1
intrapopulation / PNS / forward selection / mate choice = black / 27.4
intrapopulation / PNS / forward selection / mate choice = orange*+black / 13.8
interpopulation / PNM / backward selection / mate choice = PC1* + PC2 / 95.2
interpopulation / PNM / backward selection / mate choice = PC1** / 93.4
interpopulation / PNP / forced backward selection / mate choice = PC1• + PC2* / 119.6
interpopulation / PNS / backward selection / mate choice = PC1* + PC2 / 127.4
interpopulation / PNS / backward selection / mate choice = PC1* / 126
Significant variables are highlighted in bold, and significant levels are indicated with •<0.1, *P < 0.05, **P < 0.01. The colors correspond to the scaled color scores.
Table S2.
Summary of the most likely models. Given are in the columns from left to right the specific test (intrapopulation sexual selection or interpopulation assortative mating), the population of the females that were tested in the mate choice experiments, the linear model selection approach, the residual deviance and associated degrees of freedom (d.f.) and the explanatory variables included in the most likely model (i.e. scaled colorscores or principal components (PC)). For each explanatory variable, we reported the estimated coefficient, p- and z-value. Empty cells indicate that a particular variable was not included in the most likely model (see table S1). For the population of Python we also included the “forced” backward selection models in the table (deletion of one of the four explanatory variable), since in the most likely model individual parameters for each explanatory variable could not be uniquely estimated. For the population of PNS we only included the forward selected most likely model, since the backward selected model would not converge and thus no parameters could be estimated (see Methods and Results for details).
Test / Mate choice experiment / Model selection / Dev. / d.f. / red / orange / magenta / yellow / blue / blackintra-population / PNM / backward selection / 16.5; 20 / 3.714; 2.41; 0.016 / 4.227; 2.49; 0.013 / 2.403; 1.30; 0.193
intra-population / PNP / backward selection / 13.5; 33 / 214.73; 0.145; 0.885 / -323.73; -0.22; 0.828 / 225.98 ;0.16; 0.908 / 57.37; 0.10; 0.919
intra-population / PNP / forced backward selection / 38.7; 34 / -2.306; -2.40; 0.016 / 3.880; 2.76; 0.006 / 0.752; 1.70; 0.090
intra-population / PNP / forced backward selection / 30.2; 34 / 19.5; 3.61; <0.001 / 62.30; 4.17; <0.001 / 7.142; 3.11; 0.002
intra-population / PNP / forced backward selection / 31.2; 34 / 7.170; 1.36; 0.173 / -8.348; -1.96; 0.050 / 6.111; 2.17; 0.030
intra-population / PNP / forced backward selection / 51.6; 34 / 1.010; 1.41; 0.159 / -0.441; 0.13; 0.899 / 0.042; -0.65; 0.518
intra-population / PNS / forward selection / 3.8; 9 / -22.024; -2.25; 0.025 / 7.715; 1.35; 0.178
Test / Mate choice experiment / Model selection / Dev. / d.f. / PC1 / PC2
inter-populaiton / PNM / backward selection / 85.4; 104 / -0.043; -3.301; <0.001
inter-populaiton / PNP / forced backward selection / 109.6; 133 / 0.044; 1.80; 0.073 / -0.102; -3.29; 0.001
inter-populaiton / PNS / backward selection / 118; 134 / -0.026; -2.49; 0.013
Table 3.
Loadings of each colorscore on the PC-axes 1 and 2 from the population
specific PCA’s of Makobe, Python and Senga.
Population / PNM / PNP / PNSPC-axis / PC1 / PC2 / PC1 / PC2 / PC1 / PC2
redscore / -0.362 / 0.735 / -0.422 / -0.623 / -0.335 / 0.875
orangescore / -0.095 / 0.269 / -0.024 / -0.388 / -0.07 / 0.17
yellowscore / 0.048 / 0.431 / 0.084 / -0.547 / 0.062 / 0.183
bluescore / 0.128 / -0.035 / 0.16 / 0.186 / 0.119 / -0.116
magentascore / 0.138 / -0.284 / 0.074 / 0.184 / 0.081 / -0.192
blackscore / -0.907 / -0.347 / -0.885 / 0.305 / -0.926 / -0.349
Table S4
The number of replicates for each population mate choice experiment (PNP, PNS, PNM). Followed by the number of spawnings in each replicate, the clutch size for each spawning, the total number of fry genotyped, the percentage of fry genotyped, the male identification name, the assigned sire population and the number of mismatches to assign a fry to a sire (0-2 mismatches). The PNM population was tested at two different time periods and thus the abbreviation PNM_1 depicts the first experimental period (from October to December 2008) and PNM_2 the second experimental period (from January to March 2012).
Assignment withPopulation of female / Replicate / Spawning / Clutch size / Number of fry genotyped / % of fry genotyped / Male ID / Sire population / 0 mis-match / 1 mis-match / 2 mis-match
PNP / 1 / 1 / 23 / 9 / 39 / PNPm09 / PNP / 4 / 3 / 2
PNP / 1 / 2 / 13 / 5 / 38 / PNPm09 / PNP / 1 / 4 / 0
PNP / 2 / 1 / 27 / 7 / 26 / PNPm12 / PNP / 5 / 2 / 0
PNP / 2 / 2 / 42 / 7 / 17 / PNPm12 / PNP / 2 / 5 / 0
PNP / 2 / 3 / 17 / 12 / 71 / PNPm12 / PNP / 7 / 2 / 3
PNP / 3 / 1 / 15 / 10 / 67 / PNP16 / PNP / 5 / 5 / 0
PNP / 3 / 2 / 38 / 10 / 26 / PNMm10 / PNM / 9 / 1 / 0
PNP / 3 / 3 / 44 / 7 / 16 / PNPm16 / PNP / 2 / 4 / 1
PNP / 4 / 1 / 41 / 6 / 15 / PNPm17 / PNP / 6 / 0 / 0
PNP / 4 / 2 / 29 / 9 / 31 / PNPm17 / PNP / 7 / 2 / 0
PNP / 4 / 3 / 15 / 7 / 47 / PNPm17 / PNP / 3 / 3 / 1
PNP / 5 / 1 / 54 / 6 / 11 / PNPm09 / PNP / 0 / 4 / 2
PNP / 5 / 2 / 24 / 11 / 46 / PNPm09 / PNP / 11 / 0 / 0
PNP / 5 / 3 / 7 / 5 / 71 / PNPm09 / PNP / 4 / 1 / 0
PNP / 6 / 1 / 20 / 9 / 45 / PNPm19 / PNP / 7 / 2 / 0
PNP / 6 / 2 / 44 / 7 / 16 / PNMm08 / PNM / 2 / 4 / 1
PNP / 6 / 3 / 40 / 8 / 20 / PNPm19 / PNP / 4 / 3 / 1
PNP / 7 / 1 / 40 / 9 / 23 / PNPm20 / PNP / 9 / 0 / 0
PNP / 7 / 2 / 29 / 7 / 24 / PNPm21 / PNP / 6 / 1 / 0
PNP / 7 / 3 / 55 / 8 / 15 / PNPm20 / PNP / 6 / 2 / 0