Brains vs Brawn: Relative brain size is sexually dimorphic amongst weapon bearing ruminants

Here, we investigate the relationship between relative brain size and sexual weapons in ruminants. In most cases, sexual weaponry is heavily male-biased, and costs resulting from growing, maintaining, or wielding weapons will be suffered primarily by males. We used comparative phylogenetic analyses to test whether increased investment in sexual weapon size (tusks, antlers, and horns) across four families (Tragulidae, Moschidae, Cervidae, and Bovidae) was associated with decrease in relative brain size, and whether the difference in weapon investment relative to conspecific females led to sexual differences in relative brain size. We found no relationship between relative brain size and relative weapon size within males or females, but when we compared males directly to conspecific females, we found that as males possessed larger weaponry, they had smaller brain sizes, regardless of weapon type. Our finding suggest male investment in some types of elaborate weapons could be related to male reduction in larger brains.


47
Natural selection favoring greater cognitive ability is hypothesized to explain the 48 evolution of large brain sizes in many bird and mammal species (Eisenberg & Wilson, 1978;

54
A recent study in mammals showed that significant investments in morphological antipredator 55 defenses (e.g., spines/quills, dermal armor, noxious sprays) were associated with reductions in 56 relative brain size, suggesting that selection favoring costly morphological defenses can 57 overwhelm selection favoring advanced cognitive abilities, especially in dangerously exposed 58 environments (Stankowich & Romero, 2017). Given that intrasexual selection strongly favors 59 elaborate sexual weapons in male ruminant mammals (tusks, horns, antlers) and these can be relationship with relative brain size between males and females of the same species (i.e., when 64 males evolve to invest more in their weapon, does their brain size decrease relative to females of mammals are particularly well studied and vary considerably in size, shape, weight, growth  Nur & Hasson, 1984), which suggests that 93 investment in sexual weapons in these species could limit relative investment in other growing 94 structures.

95
Because sexual weapons in the majority of ungulates are male biased, we hypothesized a 96 negative relationship between weapons and brains, where increased investment in weaponry 97 leads to decreased investment in brain size. We predicted this effect within each sex, but we also 98 predicted that as sexual dimorphism in relative weapon investment increases in a species (i.e., as 99 males evolve larger weapons relative to females of their species), male brain size will decrease 100 relative to females of the same species. We tested for relationships between sexual weaponry and 101 relative brain size in ruminant artiodactyls by measuring weapon length (canine, antler, or horn), 102 skull length, and endocranial volume from male and female skulls of 8 tusk-bearing species 103 across three families (Tragulidae, Moschidae, and Cervidae), 13 antler-bearing cervid species, 104 and 11 horn-bearing bovid species (Fig. 1A). We calculated sex-specific measures of relative 105 brain and weapon size and used comparative phylogenetic analyses to test our predictions.

108
In this study, we tested the effects of sexual weaponry on relative brain size in ungulates History (FMNH). We took measurements on both adult male and female specimens, and only 115 included species in our analyses where we had complete measurements from at least three 116 individuals of each sex (one exception, Rangifer tarandus, only 2 females measured). While our 117 final sample was 29 species, we want to note how difficult it is to find at least 3-5 fully intact 118 male and 3-5 fully intact female skulls (that include at least one weapon and a cranium that isn't 119 broken to measure volume) of ungulate species in natural history museums, and the time it takes 120 to take these measurements. We feel that the fact that we were able to detect a significant effect 121 despite having only 29 species (and fewer in the separate weapon tests) suggests a strong 122 negative relationship and a more conservative approach. 123 We collected the following cranial measurements. Skull length (mm) was measured from    (Table 1) 161 Next, we used the resulting corrected ẞ (slope) and b (intercept) estimates to calculate 162 the predicted brain masses and weapon lengths for each individual specimen based on their 163 individual skull lengths: BrMi(predicted) = 10 b(BrMvsSkL) × SkLi ẞ(BrMvsSkL) ; WLi(predicted) = 10 b(WLvsSkV) 164 × SkVi ẞ(WLvsSkV) . EQi for each individual skull was calculated as BrMi(measured)/BrMi(predicted), 165 where an EQ above 1.0 would represent a relatively large brain and an EQ below 1.0 would be a 166 relatively small brain. Similarly, WQi for each individual skull would be calculated as 167 WLi(measured)/WLi(predicted) and be interpreted the same way relative to a value of 1.0. Antler WQ 168 was automatically set to zero for almost all female cervids, with the exception for antlered 169 female caribou (Rangifer tarandus). 170 We then calculated the average EQ and WQ for the male and female specimens for each 171 species, resulting in average EQ♂, WQ♂, EQ♀, and WQ♀ for each species based on either skull 172 length or body mass (8 total measures for each species). Next, we calculated the difference 173 between EQ♂ and EQ♀ (ΔEQ) and the difference between WQ♂ and WQ♀ (ΔWQ) for each 174 species. A result of ΔEQ below 0 indicates that females have relatively larger brains than males 175 in those species. Since females of antlered species almost exclusively had WQ♀=0, ΔWQ = WQ♂ 176 with the exception for antlered female caribou (Rangifer tarandus). We used these species 177 averages to determine whether there is a sexually dimorphic relationship between males and 178 females in relative brain investment, and whether males suffer a physiological trade-off between 179 weapon length and relative brain investment.   (Table 3). In addition, we ran phylogenetically corrected tests for 197 interaction effects on WQ or ΔWQ. All additional results can be found in our Online 198 Supplement. We set our significance level at  = 0.05 and calculated phylogenetic signal for 199 each test using maximum-likelihood estimations of lambda () derived from the PGLS tests.

205
Overall, weapon size scales hyperallometrically with body size (as estimated by skull 206 size) suggesting that as individuals (male biased) grow larger, they grow disproportionately 207 larger weapons (Table 2; Fig. S3), similar to findings in other studies (Gould, 1974 between relative investment in brain size and relative investment in weapon size. 213 We calculated sex-specific measures of relative investments in brain size 214 (encephalization quotient: EQ) and weapon size (weapons quotient: WQ) by correcting for body 215 size (as estimated by skull size): EQ and WQ scores above 1 indicate greater than expected 216 investment in these structures and scores below 1 indicate smaller investments relative to body 217 size. For males, we did not find any effect of investment in weaponry on brain size (Table 3; Fig.   218 1B), although male antlered species had significantly greater investment in their brains than 219 tusked and horned species (p < 0.05; Figure 1B; Table 3). We did not find any effect of 220 investment in weaponry on brain size in females (Table 3; Fig. 1C), although, again, female 221 antlered species had significantly greater investment in their brains than tusked species (p < 222 0.001). 223 We found a significant negative relationship between the degrees of sexual dimorphism 224 in relative brain size (EQ) and relative weapon size (WQ; Table 3, Fig. 1D), whereas males 225 evolve to invest proportionally more in weapons than females of their species, they evolve to 226 invest proportionally less in brain size (p = 0.014). This result supports that investment in relative brain size is sexually dimorphic and likely influenced by the presence of exaggerated 228 sexual weapons in these male ungulates. We also found that the difference between male and 229 female relative brain size investment was greater in antlered species than horned species (p = 230 0.049; Figure 1D; Table 3) 231 Phylogenetic signal in the analyses was either completely absent ( = 0.000) or very 232 strong ( near or equal to 1; Tables 2 & 3, S1), suggesting great variation in the degree to which 233 shared ancestry explains variation in relative brain and weapon size. We did not find any 234 evidence of an interaction effect of WQ on EQ in any analyses, so the interaction term was 235 dropped from all final models (Table S1). In addition, we ran separate analyses amongst each 236 weapon type individually (Horns=11, Antlers=13, Tusks=8). For all groups, we found similar 237 insignificant results between male WQ and male EQ, except amongst tusked species, we found a 238 significant negative relationship between male WQ and EQ and female WQ and EQ. Lastly, for 239 our sexual dimorphic analysis, we found similar results in our tusked and horned groups, with 240 significant negative relationships between M-F EQ and M-F WQ. However, we did not find any 241 relationship amongst our antlered group. All additional analyses can be found in our 242 supplemental data (Table S2; Figure S6; S7).

244
Our data support the hypothesis that increased investment into male weapon size is 245 associated with a sexual dimorphic investment in relative brain size. Across twenty-nine species 246 and three weapon types (horns, tusks, antlers), as males evolved to invest relatively more than 247 conspecific females in building larger weapons, they invested relatively less than conspecific strongly on males to invest in progressively larger weapons that it creates inequity in the brain 258 sizes of males and females, with brain size of males decreasing relative to conspecific females in 259 species with the largest weapons (Fig. 1C). Since EQ has sometimes been used as a rough but when compared to females, who lack sexual weapons, relative brain size is larger in bearing species. Bovids are unusual among ungulates because females of many species also 273 develop sizeable horns. Bovid females use their horns either for defense against predators or to 274 guard territories against conspecifics (Stankowich & Caro, 2009), so females that invest heavily 275 in horns may also face tradeoffs with relative brain size. We found patterns of sexual 276 dimorphism in relative brain size consistent with this tradeoff, as species with strong sexual 277 dimorphism in weapon size also had the largest difference between male and female brain sizes.

278
The potential difference in energy investment when developing weapons (permanent vs.  Initially, we hypothesized that, within males, as relative investment in sexual weapons 285 increased, the relative investment in brain size would drop, we found no relationship between 286 relative brain and weapon size in females or males. Although many other studies found little to  and assessment may require greater cognitive and decision-making abilities in these species, 308 strengthening selection for larger brains. In contrast, tusked species tend to be "slinkers" that 309 tend to be smaller in size, live in more closed habitats, and engage in quick slashing and stabbing 310 combat in close quarters when they meet, without much signaling. In support of this, we found 311 that tusked species, had lower EQ values in both males and females. for a larger sample size and broader taxonomic sampling. Here, we use EQs to examine the 317 relationship between relative brain size and sexual weapon size, rather than as a measure of higher cognition; though the declines in relative brain sizes we found with greater weapon sizes 319 in males relative to females may extend to cognitive effects as well. Future studies should further 320 question if males with larger sexual weapons may energetically compensate with reductions in 321 cranial thickness, musculature, fecundity, or longevity, or with significant increases in energetic 322 intake relative to females of the same species.