Evolution of neurometabolic frugality in harvester ants

11 Despite the importance of neurometabolic costs in brain size evolution, quantitative data on 12 brain metabolic rates are lacking. We measured ex vivo brain metabolic rates among species 13 of the ant genus Pogonomyrmex to differentiate the roles of sociality and body size in brain 14 evolution in a phylogenetic context. Worker body size and colony size (a proxy for social 15 complexity) vary significantly among Pogonomyrmex species and were positively correlated. 16 However, sociality was not a determinant of brain energetics. Worker body size strongly 17 affected brain metabolism: 38% of resting metabolic rate was attributable to brain 18 metabolism in species with the smallest workers, compared to 6% in species with the largest 19 workers. More derived species had strikingly lower mass-specific brain metabolic costs, 20 suggesting that increases in worker body size have selected for neurometabolic frugality 21 through reductions in brain mass-specific metabolic rate. Additionally, smaller worker body 22 sizes may require higher brain mass-specific energetic costs to achieve comparable 23 performance by absolutely smaller brains. Our study shows that the social brain hypothesis 24 does not explain patterns of brain size in Pogonomyrmex, but body size and evolutionary 25 history strongly influence brain evolution in regard to both size and metabolic cost. 26 27

28 Understanding the selective forces that drive brain evolution and its macroevolutionary pat-29 terns remains one of the most intriguing unresolved problems in biology 1–6. Body size 30 strongly affects brain size; smaller bodies have relatively larger brains (Haller’s Rule7–10), but 31 the mechanisms linking body-size related patterns to brain morphology and physiology re-32 main unclear6,9,11. In terms of social and ecological influences on brain evolution, theories hy-33 pothesize that the behavioural and cognitive demands of group living12,13 or diet and its influ-34 ences on foraging14,15 are major selective factors of brain size and mosaicism. Collaterally, 35 the expensive tissue hypothesis16 emphasizes the unusually high metabolic expense of the 36 brain and posits that trade-offs between costly neural and digestive tissue permitted the evo-37 lution of larger brain sizes. Furthermore, an adipose-brain trade-off may also explain brain 38 size evolution, with fat stores and encephalization forming alternative insurance policies 39 against starvation17. Support for these theories varies and debate continues17–19, most often 40 focusing on vertebrates1,20, particularly humans and other primates21 (and references therein). 41 While energy costs are important components of many of these theories, few studies have 42 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 2 measured brain metabolic rates (MR), and no studies have examined directional evolution of 43 the size and MR of brains, and how these interact with body size or social evolution. 44 There is considerable evidence that larger animals have absolutely larger but 45 relatively smaller brains, and some evidence that brains of larger animals use less energy per 46 gram. Brain size scales hypometrically in mammals8,22,23, birds23–25, reptiles26, amphibians27, 47 fish28, arachnids29, and insects30,31, including bees32–34 and ants35,36. Brain MR has also been 48 shown to scale hypometrically in mammals and insects11,37,38. The causes of the hypometric 49 scaling of brain size and metabolism are hypothesized to be related to constraints on whole-50 body MR8 and/or selection on life history parameters, such as the need for smaller animals to 51 maintain a behavioural capacity comparable to that of larger animals39–41. There may also be 52 selection for energy efficiency and a longer lifespan in larger animals39,41,42. Larger brains 53 may be more cognitively capable43–45, but body and brain size effects on behavioural 54 capacities are inconsistent. Generally, small-bodied species do not have limited behavioural 55 repertoires compared to larger-bodied species39,46, and the need for smaller individuals and 56 species to compete, behaviorally and cognitively, with larger competitors41 may explain 57 hypometric brain scaling. The brains of the smallest insects displace somata and lose nuceli 58 to preserve axon number, increasing relative brain mass allocated to energetically costly 59 neurons, which may generate higher brain mass-specific metabolic rate (MSMR) in smaller 60 individuals47–50. However, no study has examined the scaling of brain MR in a phylogenetic 61 context, therefore there are no data on the extent to which variation in brain metabolism is 62 associated with relatedness, independent of body and brain size variation, or whether there 63 are directional evolutionary tendencies in neurometabolism. 64 Eusocial insects have emerged as important models for analyses of brain evolution in 65 animal societies due to their diversity and size variation31,36,51–55, and here we test for effects 66 of body size, social complexity, and phylogeny on brain size and MR in harvester ants, Pogo-67 nomyrmex. The application of the social brain hypothesis to eusocial insects is controver-68 sial55–57. Worker sterility within eusocial species precludes the evolution of individual strate-69 gies of reproductive competition, demanding coalition formation and socio-cognitive skills, 70 which require increased brain tissue and a high processing capability. Eusocial insect workers 71 and groups exhibit individual and collective cognition, respectively58, but worker cognitive 72 abilities do not usually directly enhance worker fitness. In eusocial species with larger colony 73 sizes, a proxy for social complexity due to increased social interactions35,52,59–62, individual 74 behavioural/cognitive load may decrease due to task specialisation and division of labour61. 75 This could result in smaller, less energy intensive brains associated with increasing social 76 complexity51,63. Here, empirical data are equivocal21,55. Comparisons of the evolutionary neu-77 robiology of ants suggest that striking differences in colony size and degree of social com-78 plexity influence brain size and its energetic expense64. In some ants, species that form larger 79 colonies have evolved workers with relatively larger brains35. Sociality is also associated with 80 larger brains in some bees33,34,65,66 but the broader pattern in this clade also indicates a signifi-81 cant role for diet and life history32,67. Conversely, in monomorphic fungus-growing ants68 and 82 vespid wasps52, increased social complexity is associated, respectively, with decreased rela-83 tive size of brains and mushroom bodies, the brain compartments specialized for learning and 84 memory67. Together these results suggest that the effects of social complexity on social insect 85 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 3 brains are complex and likely affected by ecology, colony size, social system, and eusocial 86 insect clade53,56. 87 Phylogenetic history also appears to play important roles in brain evolution. Different 88 clades often have different absolute and relative brain sizes6,9,32,36, and can have different 89 relationships between brain size and body size6. The Marsh-Lartet rule, which suggests an 90 evolutionary trend of increasing relative brain mass in mammals, was recently supported for 91 three mammalian orders including primates6,69. In bees, phylogenetic analysis revealed that 92 brain size evolution is linked to evolutionary changes in voltinism and host specialization32. 93 Understanding the directional evolution of brain size and MR, and how tightly these are 94 constrained by body size, are important questions for evolutionary biology. Unfortunately, we 95 currently lack studies of directional evolution and phylogenetic patterns for brain metabolism 96 in any animal clade. 97 To address major gaps in our understanding of the evolution of neurometabolism, 98 brain size and social system, we measured brain masses, whole-body MRs, and intact brain 99 MRs ex vivo from workers of eleven species of seed-harvesting ants, Pogonomyrmex70. Ants 100 in general71, and Pogonomyrmex in particular, provide an excellent model to address 101 questions at the intersection of neurobiology, socioecology, and metabolism. A larger body 102 size is known to be associated with a relatively smaller brain within leafcutter ant subcastes36, 103 within Cataglyphis species35, across 70 species of ants36, and more generally in insects9. 104 Pogonomyrmex have significant variation in body mass, colony size, and foraging ecology. 105 The species studied here are all primarily granivorous and reproduce annually, enabling tests 106 for effects of body size, colony size and phylogeny, without obvious confounding variables 107 of life history or diet (Table 1). All 11 species inhabit grasslands of the southwestern United 108 States and robust phylogenies of the genus exist72, along with assessments of colony size 109 (Table 1). We determined the effects of body size and social complexity on brain and body 110 MRs in a phylogenetic context to gain insight into possible evolutionary trends. We ask: how 111 is the evolution of Pogonomyrmex linked with changes in body size, brain size, brain and 112 whole-body MRs, and social complexity? 113 114

115 i. Phylogenetic signal 116 There was a significant phylogenetic signal for brain MSMR (Fig. 1; Pagel’s λ = 1.00, p < 117 0.01), body MSMR (Fig. 1; Pagel’s λ = 1.00, p < 0.01), brain MR (Fig. 1; Pagel’s λ = 0.98, p 118 < 0.01), and colony size (Fig. 1; Pagel’s λ = 1.00, p < 0.01). For brain mass-specific MR and 119 body MSMR there was a trend for more derived species to have lower rates (Fig. 1). In 120 contrast, for brain MR and colony size, there was a trend for more derived species to have 121 higher values (Fig. 1). There was no significant phylogenetic signal for body mass (Pagel’s λ 122 = 0.55, p > 0.20), brain mass (Pagel’s λ = 0.65, p > 0.20) or whole-body MR (Pagel’s λ < 123 0.01, p = 1.00). As such, subsequent analyses using body mass, whole body MR, or brain 124 mass were conducted without any phylogenetic correction, but any analysis of brain MR, 125 mass-specific MR or colony size accounted for phylogeny. 126 127 ii. Body size and colony size 128 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 4 Mean worker mass increased with increasing colony size (Fig. 1; PGLS; t9,11 = 2.73, p < 129 0.03). 130 131 iii. Brain size scaling and relationship with colony size 132 There was a significant relationship between brain mass and body mass (Fig. 3A; linear 133 mixed effects model; t9,252 = 15.73, p < 0.001). Brain mass scaled hypometrically, with an 134 allometric slope of 0.54 ± 0.03 (for this and subsequent regression equations, the standard 135 error is provided). There was a significant relationship between brain mass and colony size 136 (Fig. 3B; PGLS; t9,11 = 3.07, p < 0.02): worker brain mass increased with increasing colony 137 size. There was no significant relationship between relative brain mass (brain mass per unit 138 body mass) and colony size (Fig. 3C; PGLS; t9,11 = 0.52, p = 0.62). 139 140 iv. Whole body resting metabolic scaling and relationship with colony size 141 There was a significant relationship between log10 whole body MR and log10 body mass (Fig 142 4A; linear mixed effects model; t37,252 = 3.10, p < 0.01). Whole body resting MR scaled 143 hypometrically; the allometric slope was 0.35 ± 0.11. There was a significant relationship 144 between whole body MR and colony size (Fig. 4B; PGLS; t9,11 = 2.44, p < 0.04). Whole body 145 MR increases with increasing colony size. However, the relationship between whole body 146 MR and colony size becomes nonsignificant if body size is taken into account. When colony 147 size and body mass were both included as predictors of body MR in a multivariate model, 148 body mass emerged as a significant factor (PGLS; t1,11 = 4.09, p < 0.01), and colony size does 149 not (PGLS; t1,11 = 0.95, p = 0.37). There was no significant relationship between whole body 150 MSMR and colony size (Fig. 4C; PGLS; t9,11 = 0.35, p = 0.73). Furthermore, there was no 151 significant relationship between body-mass-corrected MR (residuals of whole-body MR 152 regressed against body mass) and colony size (Fig. S2A; PGLS; F1,11 = 1.13, p = 0.29). 153 154 v. Brain metabolic rate scaling and relationship with colony size 155 There was no significant relationship between brain MR and colony size (Fig. 5A; PGLS; t9,11 156 = 0.77, p = 0.46) and there was no significant linear relationship between brain MR and brain 157 mass (PGLS; t9,11 = 0.84, p = 0.42). There was, however, a significant non-linear relationship 158 between brain MR and brain mass (Fig. 5B; GAM with thin plate spline and phylogenetic 159 penalty; F1,11 = 3.21, p < 0.01). Brain MR increased with increasing brain mass up to a value 160 of 0.148 mg and then decreased slightly. There was a significant relationship between brain 161 MSMR and colony size (Fig. 5C; GAM with thin plate spline and phylogenetic penalty; F1,11 162 = 7.06, p < 0.03). There was also a significant linear fit between brain MR and colony size 163 (PGLS; t9,11 = 2.33, p < 0.05), but the non-linear model was selected due to having a lower 164 AIC score (GAM = 70.89, PGLS = 75.02). Brain MSMR decreased with increasing colony 165 size. There was also a highly significant relationship between brain MSMR and body mass 166 (Fig. 5D; PGLS; t1,11 = 5.70, p < 0.001). As body mass increased, the relative energetic cost 167 per gram of brain decreased; brains from smaller individuals used more energy per unit mass 168 of brain than larger individuals. When colony size and body mass were both included as 169 predictors of brain MSMR in a multivariate model, body mass emerged as a significant factor 170 (PGLS; t1,11 = 5.70, p < 0.001), and colony size does not (PGLS; t1,11 = 0.68, p = 0.52). There 171 was no significant interaction term (PGLS; t1,11 = 0.90, p = 0.40). Furthermore, there was no 172 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 5 significant relationship between body-mass-corrected brain MSMR (residuals of brain 173 MSMR regressed against body mass) and colony size (Fig. S2B; PGLS; F1,11 = 0.33, p = 174 0.58). 175 176

177 Phylogeny and body size had strong effects on brain and body energetics, but colony size, a 178 proxy for social complexity, did not. There was a general trend for more derived species to 179 have lower brain and body MSMR (Fig. 1). Body size did not have a phylogenetic signature, 180 but was strongly correlated with brain MSMR, with larger-bodied species having lower 181 MSMRs (Fig. 5D). Higher brain MSMRs were also correlated with increasing social 182 complexity, but this relationship was non-significant when body size was accounted for (Fig. 183 S2B). Brain MR increased with brain size only up to a brain mass of 0.148 mg (Fig. 5B). 184 Species with larger brains (e.g. P. barbatus, P. rugosus) had equivalent brain MRs to those 185 with significantly smaller brains (e.g. P. apache, P. badius, P. occidentalis). Colony size was 186 strongly correlated with body size; larger colonies had larger workers (Fig. 1). There were 187 some significant relationships between colony size and energetics: workers from larger 188 colonies showed higher body MRs. However, this relationship was not independent of an 189 increase in worker body size (Fig. S2A), and colony size was not a significant predictor of 190 body MR with body mass accounted for. Therefore, body size was the more important factor 191 for determining body and brain energetics. We identified strong directional evolutionary 192 tendencies for these traits, with brain and body MSMR being lower in more derived species, 193 indicating a directional evolutionary trend towards energetic frugality. 194 Though body size is the most important factor for determining energetics, how this 195 trait affects the association of larger workers with larger colonies remains unclear. Worker 196 size is inconsistently associated with colony size in social insects; some ant species (e.g. 197 Solenopsis invicta, Linepithema humile) with the most populous colonies have relatively 198 small workers and others are exceptionally polymorphic. However, across more than 100 ant 199 species, worker mass, queen mass and colony mass are strongly positively correlated74. The 200 positive association between worker size and colony size in our sampled Pogonomyrmex is 201 therefore similar to broad, but not universal, trends across ants. The relationship between 202 worker body size, foraging productivity, and energy efficiency is important but unclear, and 203 may not reflect vertebrate patterns due to the impacts of larger colony sizes on worker 204 resource acquisition. Across multiple vertebrate taxa, species with larger body sizes use more 205 resources per individual, but as individual abundances and generation times scale inversely to 206 individual size in vertebrates, large- and small-bodied species tend to utilize similar amounts 207 of energy per year (the equal fitness paradigm73,74). There appear to be advantages to larger 208 colony sizes. Larger colonies, associated with larger body-sizes in Pogonomyrmex, may 209 control clumped resources75,76 (though see77) and avoid conflict with neighbouring colonies78, 210 but it is unknown how resource acquisition compares across small and large colonies and 211 whether they also conform to the equal fitness paradigm. Within Pogonomyrmex, group or 212 trunk-trail foraging is restricted to larger bodied species and larger colonies79, therefore it is 213 possible that the energetic benefits brought by the evolution of larger workers with access to 214 differential foraging strategies can support larger colony sizes, though in other seed-215 harvesting species smaller individuals are associated with group foraging strategies79. It is 216 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 6 therefore more likely that group foraging allows for resource domination in a typically 217 stochastic environment. Individual foraging trips, irrespective of foraging strategy, have 218 minimal impact on worker energy budgets80,81 but group foraging strategies can ease the 219 exploitation of abundant resources, when they appear79. Pogonomyrmex colonies can 220 experience significant interspecific food resource competition76 thus influencing the 221 evolution of worker size through character displacement, reducing competition by niche 222 segregation, or selecting for differential foraging strategy82,83. Colonies of P. barbatus, a 223 species with the largest workers, have lower fitness when neighbouring colony density is 224 higher84. Resource competition may favour the evolution of traits such as territoriality and 225 trunk-trail foraging, enabling resource control and, collaterally, larger workers and colony 226 sizes. Comparative data on foraging success, colony fitness, generation time, abundance, and 227 fitness of these species in the field are lacking, and the hypothesis that large worker size, 228 social complexity, and ecological dominance are linked in Pogonomyrmex requires additional 229 research, especially to determine the causality among the evolution of colony size, body size, 230 foraging strategy, and their metabolic consequences. 231 We did not find a significant effect of colony size on whole-body MSMR, perhaps 232 suggesting that larger colonies are not more metabolically efficient in our sample of 233 Pogonomyrmex species. However, it is important to note that colony MRs cannot be 234 predicted from the resting MRs of individual workers85. Large colony sizes in social insects 235 correlate with lower colony-level MSMRs86 as well as increased division of labour in P. 236 californicus61,87. Worker task diversity increases as colonies grow87 with older, larger 237 colonies having workers that tend to specialise on specific tasks, rather than engaging in a 238 wider range of tasks61,85, in addition to lower activity rates85,88. Such behavioural 239 specializations could allow larger colonies to reduce mass-specific energy use. There were 240 extremely high effect sizes of phylogeny on brain MR, brain MSMR, and whole-body 241 MSMR (Fig. 1). We did not find a phylogenetic signal for brain mass, diverging from 242 evolutionary trends in mammals, where directional evolution for increasing relative brain 243 mass has been reported within some mammalian orders, including primates6,69. 244 Although our analyses were performed on a single genus, the body size range across 245 the Pogonomyrmex species in our study is comparable to the range exhibited within 246 mammalian orders, allowing comparisons of scaling across these diverse clades. The distinct 247 life histories and reproductive traits of mammals and eusocial insects may explain the 248 differential neurometabolic response to increasing social complexity. Our data indicate that 249 the social brain hypothesis may have limitations when applied to the brain energetics of 250 eusocial insects51. We find concordance with other investigations into the association of 251 sociality and neurometabolism in ants. Kamhi et al.64 report that Australasian weaver ants, 252 Oecophylla smaragdina, a paradigm of insect social complexity, have larger brains and 253 greater investment in the mushroom bodies but significantly lower cytochrome oxidase 254 (COX) activity in the mushroom body medial and lateral calyces relative to those of the 255 socially basic, small-colony sister species Formica subsericea. The mushroom bodies are a 256 brain compartment dedicated to higher-order processing, integration, learning, and memory 257 and thus likely provide neural support for complex social behaviour in ants 63. COX is a 258 proxy for ATP usage; mass-specific brain energy expenditure may decrease as social 259 complexity increases. Our data support this pattern by directly measuring brain energetics. 260 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 7 We find the same relationship in Pogonomyrmex: brain MSMR decreases with increasing 261 social complexity, rather than increasing as would be expected from social brain theory. 262 Phylogeny is more important for both brain and whole body MSMR, indicative of 263 selection towards energetic frugality; the highest values occurred in the most basal clade (P. 264 imberbiculus and P. pima), followed by the next most basal species (P. huachucanus). The 265 lowest values occurred in the most derived clades, strongly suggesting an evolutionary trend 266 toward lower brain and whole- body MSMRs. Absolute brain MR showed the opposite 267 pattern: the lowest values were found in the more basal clades. Interestingly, brain and body 268 mass showed no phylogenetic signal, suggesting that the above patterns are not simply driven 269 by brain and body mass, and that mass is more evolutionarily labile than metabolic rate. 270 Phylogenetic signal for body mass-related traits and MRs has been found to be minimal in 271 some insects (e.g. stingless bees89, scarab beetles90), whereas others show strong effects (e.g. 272 size-related variables of bees and moths89,91,92). Our sampling of 11 species in Pogonomyrmex 273 should be extended to confirm these patterns, as we measured approximately a third of the 274 known North American species93. 275 Larger-bodied ant workers have remarkably lower energy usage per unit mass of 276 brain. The smallest-bodied species have 10x higher brain MSMR than the largest-bodied 277 species in Pogonomyrmex. Due to their relatively large brains and higher brain MSMRs, 278 brains account for a greater proportion of the metabolic budget in smaller-bodied species 279 (38% in P. huachuncanus, 29% in P. desertorum and P. imberbiculus) but only 6% in the 280 largest-bodied species (P. rugosus). Even though brains account for only 2% of whole-body 281 mass, this scaling of brain MR explains ~25% of the hypometric scaling of whole-body MRs 282 across the clade. This further demonstrates that body size is of high significance for 283 determining neurometabolic costs and indicates that a small body size is associated with both 284 relatively larger and more energetically active brains. The high energetic cost incurred by 285 brains may provide a selective pressure to reduce relative brain MRs, as observed in the more 286 derived Pogonomyrmex species. 287 The strong hypometric scaling of brain MR suggests either selection to reduce brain 288 MRs in larger-bodied species, selection for high brain MRs in smaller-bodied species, or 289 both. Limits to whole-body MR, such as increasingly constrained oxygen or nutrient delivery 290 in larger brains, is a possible cause of hypometric brain MR scaling8. However, we currently 291 lack studies of the scaling of oxygen and substrate supply in ants as have been performed for 292 some other groups of insects94,95, and no studies have yet addressed this issue for any brains. 293 Brains, and neural tissue in general, are often cited as being a significant constituent of 294 overall energy budget11,16. As such, large-bodied species with the absolute largest brain 295 masses might gain the most benefit from reducing brain MSMRs11,96. Larger-bodied species 296 generally have longer lifespans (reviewed by42), and there is evidence that this is also true in 297 ants97,98. Higher MSMRs are often associated with higher rates of reactive oxygen species 298 production, which can damage tissues and shorten lifespan, therefore it is possible that the 299 lower MSMRs of larger-bodied species arise ultimately from selection to reduce metabolic-300 linked tissue damage and extend lifespan99,100. 301 Alternatively, or additionally, species with smaller workers may experience greater 302 selection to increase their brain MRs. Small harvester ant workers carry out the same tasks as 303 larger-bodied species (e.g. forage for seeds, defend territory and/or food sources, navigate, 304 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 8 build and maintain the nest, care for the queen and brood) with brains and sensory organs that 305 are absolutely smaller; this may require increased intensity of brain operations. Furthermore, 306 smaller-bodied seed-harvesting species are often sympatric with larger-bodied species and 307 likely compete to some extent for seeds76. Though seed sizes are positively correlated with 308 worker body size83, interspecific foraging competition76 could lead to stronger selection on 309 mass-specific brain performance in smaller-bodied species41. 310 The non-linear relationship between brain MR and brain size provides further 311 evidence that species with smaller-brained workers may be under significant selection to 312 maximise their performance, and therefore brain MRs. Brain MR did not vary between the 313 largest species as much as between the smallest species. The steep increase in brain MR with 314 brain size when brain mass < 0.15 mg indicates that there is a benefit to increasing brain MR 315 at smaller brain sizes, or brain energy consumption is limited at larger brain sizes. 316 The high brain MSMR of smaller-bodied workers may also be a function of other 317 aspects of miniaturization47. Smaller insects preserve axons over glial cells in the central 318 nervous system, likely to maintain brain function. Generating action potentials and synaptic 319 signalling are the costliest neuronal functions47–50. Therefore, the necessity to maintain brain 320 function despite miniaturization might require increasing intensity of brain use and high brain 321 MSMRs. Coupled with the strong propensity for sensory organs to also scale 322 hypometrically101–104, higher mass-specific information processing for smaller animals could 323 partially preserve cognitive and behavioural capacities53–56. This may be a proximate 324 mechanism by which smaller individuals maintain cognitive capacity to compete with larger-325 bodied sympatric species41. Additionally, smaller-bodied species are often more likely to 326 experience predation, which may also select for better sensory systems and reaction times42. 327 Further behavioural, ecological, and physiological studies will be required to determine the 328 causal factors driving body-size associated energetic patterns. 329 Although our metabolic measurements were made on intact undamaged brains, we 330 recognize that the dissection necessary for recording ex vivo brain metabolic rates has 331 physiological effects that are not well understood. Severing sensory and motor neurons of the 332 brain will likely first increase ion fluxes and therefore cause apoptosis, though neurons have 333 mechanisms to protect against this process105. The number of neurons cut, however, is likely 334 extremely small relative to the total number of brain cells. Smaller P. rugosus workers have 335 50,000-80,000 brain cells and comparable-size workers of other ant species have upwards of 336 200,000 cells105. Furthermore, the stability of brain metabolic rates over several hours 337 demonstrates that dissections are not driving a fast, progressive deterioration of brain 338 function. Data from in vitro tissues, including the brain, in fish indicate similar metabolic 339 scaling exponents as those measured from in intact animals106,107, and glucose uptake by the 340 brains of starved rats are similarly depressed in vivo and ex vivo108, suggesting that the 341 relationship among metabolic levels of brains are not negatively impacted by removal from 342 the body. Many studies have successfully utilised the Seahorse system to detect the effects of 343 toxins109, diet110 and trauma, even from partial brains that have undergone destructive 344 sampling111. Together these studies suggest that such ex vivo brain metabolism measures 345 accurately reflect in vivo interspecific patterns, opening avenues for future research. 346 Size is frequently used as a proxy for brain energetic cost, neuropil investment, and 347 behavioural and/or cognitive processing ability20. If absolute brain size is considered as the 348 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 9 index of brain investment, one might conclude that the social brain theory was supported in 349 this clade (Fig. 3B); however, incorporation of the effects of size and mass-specific metabolic 350 rate (Figs. 3C, 4C) result in the opposite conclusion. Linkages between brain size (and neuron 351 number) and cognitive ability grow weaker as greater taxonomic diversity is considered71, 352 indicating that brain size might be an inconsistent metric for behavioural capability. Our brain 353 MR data provide empirical evidence that brain size is not necessarily an accurate proxy for 354 brain energy use20,71. Pogonomyrmex rugosus workers have the largest brain in our study 355 (0.229 mg), approximately double the mass of P. occidentalis (0.114 mg). However, the 356 difference in whole-brain energy consumption (P. occidentalis: 1.76 µW; P. rugosus: 1.64 357 µW) is minimal due to the much higher brain MSMR in the smaller-bodied species. Our 358 study reveals that assessment of brain metabolism in a phylogenetic context can be a 359 powerful tool to understand patterns of brain evolution across diverse animal clades. 360 361 Declaration of interests 362 The authors declare no competing interests. 363 364 Data and code availability 365 Data and code used in this manuscript have been deposited at Open Science Foundation and will be 366 publicly available on publication. 367 368 Ethics Statement 369 The animals used in this study are not subject to any formal legal or ethical oversight. 370 However, care was taken during handling, husbandry and experimentation to minimise any 371 potential suffering. 372 373

374 i. Study species 375 Workers of eleven species (Table 1) of Pogonomyrmex ants were collected from multiple 376 colonies and locations in Arizona and California, USA between May 2021 and January 2023 377 (Supplementary Table S1). Ants were housed in small plastic enclosures and stored in an 378 environmental control room with the temperature fixed at 30°C. They were provided with 379 30% w/v sucrose solution and fresh water ad libitum and were processed for experimentation 380 within 3 weeks of collection. 381 382 ii. Sociometrics 383 We assessed the effect of social organization on brain mass and metabolism using colony size 384 as an estimator of social complexity. Colony size is considered a reliable indicator of derived 385 and advanced sociality in ants62 and other eusocial insects35,52,59,60 owing to its association 386 with the potential number of interactions among workers, whether or not the relationships 387 among workers or groups of workers are differentiated. Further evidence for colony size as a 388 robust proxy of social complexity can found in P. barbatus, where foraging dynamics are 389 modulated by social interaction112,113. The members of these colonies experience high levels 390 of social interaction and form some of the largest colonies in the genus. 391 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 10 392 iii. Whole body resting metabolic rate measurements 393 To determine the whole body metabolic of ants at rest, we used differential flow-through 394 respirometry. The respirometer was in the same temperature-controlled (30°C) room the ants 395 were stored in. We conducted all respirometry trials in the dark, with red filters applied to 396 respirometry chambers to prevent incidental light reaching the ant during preparatory lab 397 work. 398 A Flowbar-8 mass flow meter system (Sable Systems, Las Vegas, NV, USA) pumped 399 dry, CO2 free air from a gas cylinder into the respirometry chamber at a STP flow rate of 50 400 ± 1 ml min-1. Ants were placed into a chamber consisting of Bev-A-Line tube (length: 25mm, 401 diameter: 3.2mm). Output from the chamber was directed to the sample cell of a LI-7000 402 CO2/H2O Gas Analyser, which sampled at 1Hz, and was calibrated with two CO2 calibration 403 standards analysed ± 0.2 ppm. 404 We recorded a baseline measurement, without an ant, for at least one minute 405 preceding each respirometry trial. After introducing the ant to the chamber, we covered the 406 chamber with a transparent red filter and allowed the ant to adjust to the chamber until we 407 observed little to no movement. We recorded CO2 production and synced this output with ant 408 activity using a web camera (Logitech HD Pro Webcam C920, 1080p) for 30 minutes. 409 We digitized the analog output from the LI-7000 using a Sable Systems UI2 and 410 recorded once per second using ExpeData (Sable Systems, v. 1.7.2) for Windows. We 411 calculated average CO2 levels during time periods when we observed ants to be still. 412 CO2 production rates (ml h-1) were calculated using Equation 1, with FR equal to the 413 flow rate (ml h-1), and FCO2 equal to the fractional CO2 level (μmol mol-1) in the excurrent 414 air from the respirometry chamber: 415 416 VCO2 = FCO2·FR (1) 417 We did not measure the respiratory quotient of all our species. The respiratory quotient of 418 worker ants has been reported as 0.71, 0.77, 0.8, 0.91, 0.92. and 1.02114–118. We converted 419 worker CO2 production to oxygen consumption using the average respiratory quotient of 420 these studies (0.85) and then calculated metabolic rate in microwatts assuming 20.4 joules ml 421 oxygen-1 119. Ants were sacrificed by freezing after being weighed using a XPE56 XPE 422 microanalytical balance (Mettler Toledo, Columbus, OH, USA). Sample sizes for each 423 species can be found in Table S2. 424 425 iv. Brain metabolic rate measurements 426 Brain oxygen consumption rates were measured ex vivo using a Seahorse XF HS Mini 427 Analyser (Agilent, Santa Clara, California, USA) following methods outlined in120. Workers 428 were anaesthetised on ice and their fresh masses recorded using an XPE56 XPE micro-429 analytical balance (Mettler Toledo, Columbus, OH, USA). While remaining unconscious 430 from the cold, workers were decapitated. Brains were then removed intact and undamaged 431 from the head capsule under a dissecting microscope in Seahorse XF base media (Agilent, 432 Santa Clara, California, USA) supplemented with 0.01 moles per litre glucose and sodium 433 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 11 pyruvate. The Seahorse XF HS Mini Analyser has eight wells, two of which were left empty 434 as control wells, leaving six which contained intact dissected brains. Assays were conducted 435 at 32°C as this was the temperature at which the Seahorse stabilized in the lab with the heater 436 turned off and was very near the holding and respirometry temperature used. 437 Brain MRs were recorded in cycles. The Seahorse XF HS Mini mixes oxygen into the 438 media within the well and then measures the rate at which oxygen is depleted. This cycle was 439 repeated twelve times, taking approximately an hour; MRs were usually very stable over this 440 period. The rates estimated from the last three cycles were selected and the mean MR was 441 calculated (Fig. S1). The Seahorse XF HS Mini measures oxygen consumption rates 442 (picomols min-1). Insect brains are believed to catabolize a mixture of carbohydrates and fatty 443 acid metabolites121. In the absence of knowing the exact brain respiratory quotient, we 444 converted oxygen consumption into µwatts using the same conversion factor we used for 445 resting workers, 20.4 joules per ml O2 consumed. After completion of the assay, brain mass 446 was measured using a XPE56 XPE microanalytical balance (Mettler Toledo, Columbus, OH, 447 USA). Sample sizes for each species can be found in Table S2. 448 449 v. Statistical analysis 450 All statistical analyses were conducted using R version 4.2.2122. Phylogenetic signal was 451 calculated using the package ‘phytools’123, Brownian correlation structures for phylogenetic 452 generalised least squares (PGLS) regressions were calculated using the package ‘ape’124 and 453 data visualisation was conducted using the package ‘ggplot2’125 and ‘ggtree’126. Generalised 454 additive models (GAM) were implemented using the package ‘mgcv’127, phylogenetic 455 penalties were implemented using the package ‘MRFtools’128. 456 MRs and brain masses were analysed using linear models constructed using the R base 457 package122 or mixed effects models constructed using the ‘lme4’ package129. Significance of 458 terms in mixed effects models were assessed using the package ‘lmerTest’130. MRs and body 459 masses were log10 transformed to facilitate allometric analysis. 460 461 vi. Phylogenetic reconstruction 462 11 samples from the 143 sample phylogeny inferred in Graber et al.72 were used to construct 463 the phylogeny used in our analyses. UCE library prep and enrichment protocols are described 464 in Graber et al 72. in prep. 851 alignments with >95% representation of the 11 taxa were 465 concatenated into a single matrix using ‘phyluce_align_get_only_loci_with_min_taxa’. IQ- 466 TREE2 was used to infer a maximum likelihood phylogeny with 1000 bootstrap replicates 467 and the GTR+G model used for the entire alignment. 468 469 470 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 12 TABLES 471 Table 1: Means of variables from workers of each species.472 Species Body mass (mg) Brain mass (mg) Body metabolic rate (μWatts) Brain metabolic rate (μWatts) Worker number per colony Colony size

P. apache 12.35 0.15 16.09 1.65 80 Cole (1954)131 P. badius 14.12 0.14 11.52 1.72 5000 Tschinkel (2017)132 P. barbatus 14.95 0.19 19.74 1.68 12000 Gordon (1992)133 P. californicus 6.08 0.11 6.75 1.60 3250 RA Johnson, pers. obvs P. desertorum 6.03 0.10 5.66 1.64 500 Creighton (1956)134 P. huachucanus 4.29 0.08 4.11 1.60 150 Creighton (1952)135; RA Johnson, pers. obs. P. imberbiculus 2.12 0.06 4.91 1.45 125 Heinze, et al. (1992)136 P. maricopa 9.80 0.12 5.86 1.63 750 RA Johnson, pers. obs. P. occidentalis 7.91 0.11 9.20 1.76 3880 Lavigne (1969)137 P. pima 1.68 0.05 5.02 1.41 250 Johnson, et. al (2007)138 P. rugosus 15.87 0.29 29.81 1.64 8600 MacKay (1981)139 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 13 FIGURES 473 474 Figure 1: (A) Brain mass-specific and body mass-specific metabolic rates were strongly affected by phylogeny. (B & C) The basal clades of the genus had relatively high brain and body mass-specific metabolic rates, while the rugosus -barbatus clade had the lowest, suggestive of a progressive evolutionary trend toward lower brain and body mass-specific metabolic rates. (D) Absolute brain metabolic rates also had a significant phylogenetic signal; being generally lower in more basal species. (E) Mean colony sizes had a significant phylogenetic signal, being generally smaller in more basal species. Colours are not indicative of any value but included to aid readability. .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 14 475 476 Figure 2: Worker masses increase with increasing social complexity . Mean worker mass increases with increasing colony size ( worker mass = 0.000000634* colony population + 0.004862386). .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 15 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 Figure 3: (A) Brain mass scaled hypometrically with worker body mass (log10 brain mass = 0.54 * log 10 body mass +0.20) (B) Brain mass increased with increasing colonial population (brain mass = 0.00000666 * colony size + 0.08179035). (C) Relative brain mass was not significantly related to colonial population size. Figure 4: (A) Whole body metabolic rate increased in larger workers but scaled hypometrically (log10 body metabolic rate = 0.35 * log 10 body mass + 1.62). (B) Whole body metabolic rate increased with increasing colonial population ( Body MR = 0.001035 * colony size + 6.67). (C) Whole body m ass-specific metabolic rate was not significantly related to colonial population size. .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 16 508 509 510 511 512 513 514 515 516 517 518 519 520 521 Figure 5 : Phylogenetic generalized least squares analyses of independent effects on brain metabolism. (A) There was no relationship between brain metabolic rate and colonial population. (B) There is a significant relationship between brain metabolic rate and brain mass. (C) Brain mass-specific metabolic rate decreased with increasing colonial . (D) Brain mass-specific metabolic rate also declined with increasing worker body mass (Brain MSMR = -874.87 * body mass + 26.94). When analysing body mass and colony size as predictors of brain MSMR in a multivariate model, colony size ceases to be a significant predictor of brain MSRS and body mass remaining a significant factor. .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 17

522 1. Heldstab, S. A., Isler, K., Graber, S. M., Schuppli, C. & van Schaik, C. P. The 523 economics of brain size evolution in vertebrates. Current Biology 32, R697–R708 524 (2022). 525 2. Finarelli, J. A. & Flynn, J. J. Brain-size evolution and sociality in Carnivora. Proc Natl 526 Acad Sci U S A 106, 9345–9349 (2009). 527 3. Chambers, H. R., Heldstab, S. A. & O’Hara, S. J. Why big brains? A comparison of 528 models for both primate and carnivore brain size evolution. PLoS One 16, e0261185 529 (2021). 530 4. Watanabe, A., Balanoff, A. M., Gignac, P. M., Gold, M. E. L. & Norell, M. A. Novel 531 neuroanatomical integration and scaling define avian brain shape evolution and 532 development. Elife 10, (2021). 533 5. Dembitzer, J., Castiglione, S., Raia, P. & Meiri, S. Small brains predisposed Late 534 Quaternary mammals to extinction. Sci Rep 12, (2022). 535 6. Venditti, C., Baker, J. & Barton, R. A. Co-evolutionary dynamics of mammalian brain 536 and body size. Nature Ecology & Evolution 2024 8:8 8, 1534–1542 (2024). 537 7. Van Der Woude, E., Smid, H. M., Chittka, L. & Huigens, M. E. Breaking Haller’s Rule: 538 Brain-Body Size Isometry in a Minute Parasitic Wasp. Brain Behav Evol 81, 86–92 539 (2013). 540 8. Burger, J. R., George, M. A., Leadbetter, C. & Shaikh, F. The allometry of brain size in 541 mammals. J Mammal 100, 276–283 (2019). 542 9. Eberhard, W. G. & Wcislo, W. T. Grade changes in brain–body allometry: 543 Morphological and behavioural correlates of brain size in miniature spiders, insects 544 and other invertebrates. in Advances in Insect Physiology (ed. Casas, J.) vol. 40 155–545 214 (2011). 546 10. Rensch, B. Histological changes correlated with evolutionary changes of body size. 547 Evolution (N Y) 2, 218–230 (1948). 548 11. Karbowski, J. Global and regional brain metabolic scaling and its functional 549 consequences. BMC Biol 5, 18 (2007). 550 12. Dunbar, R. The social brain hypothesis. Evolutionary Anthropology: Issues, News, 551 and Reviews 6, 178–190 (1998). 552 13. Shultz, S. & Dunbar, R. I. M. Socioecological complexity in primate groups and its 553 cognitive correlates. Philosophical Transactions of the Royal Society B 377, (2022). 554 14. DeCasien, A. R., Williams, S. A. & Higham, J. P. Primate brain size is predicted by 555 diet but not sociality. Nat Ecol Evol 1, 1–7 (2017). 556 15. Rosati, A. G. Foraging cognition: Reviving the ecological intelligence hypothesis. 557 Trends Cogn Sci 21, 691–702 (2017). 558 16. Aiello, L. C. & Wheeler, P. The expensive-tissue hypothesis: The brain and the 559 digestive system in human and primate evolution. Curr Anthropol 36, 199–221 (1995). 560 17. Navarrete, A., Van Schaik, C. P. & Isler, K. Energetics and the evolution of human 561 brain size. Nature 2011 480:7375 480, 91–93 (2011). 562 18. Dunbar, R. I. M. & Shultz, S. Why are there so many explanations for primate brain 563 evolution? Philosophical Transactions of the Royal Society B: Biological Sciences 564 372, (2017). 565 19. Sobrero, R., May-Collado, L. J., Agnarsson, I. & Hernández, C. E. Expensive brains: 566 ‘Brainy’ rodents have higher metabolic rate. Front Evol Neurosci 3, 10044 (2011). 567 20. Kverková, K. et al. The evolution of brain neuron numbers in amniotes. Proc Natl 568 Acad Sci U S A 119, e2121624119 (2022). 569 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 18 21. DeSilva, J., Traniello, J., Claxton, A. & Fannin, L. When and why did human brains 570 decrease in size? A new change-point analysis and insights from brain evolution in 571 ants. Front Ecol Evol 9, 742639 (2021). 572 22. Smaers, J. B. et al. The evolution of mammalian brain size. Sci Adv 7, 9 (2021). 573 23. Tsuboi, M. et al. Breakdown of brain–body allometry and the encephalization of birds 574 and mammals. Nature Ecology & Evolution 2018 2:9 2, 1492–1500 (2018). 575 24. Guay, P. J., Weston, M. A., Symonds, M. R. E. & Glover, H. K. Brains and bravery: 576 Little evidence of a relationship between brain size and flightiness in shorebirds. 577 Austral Ecol 38, 516–522 (2013). 578 25. Nealen, P. M. & Ricklefs, R. E. Early diversification of the avian brain:body 579 relationship. J Zool 253, 391–404 (2001). 580 26. De Meester, G., Huyghe, K. & Van Damme, R. Brain size, ecology and sociality: A 581 reptilian perspective. Biological Journal of the Linnean Society 126, 381–391 (2019). 582 27. Taylor, G. M., Nol, E. & Boire, D. Brain regions and encephalization in anurans: 583 Adaptation or stability? Brain Behav Evol 45, 96–109 (1995). 584 28. Triki, Z., Aellen, M., Van Schaik, C. P. & Bshary, R. Relative brain size and cognitive 585 equivalence in fishes. Brain Behav Evol 96, 124–136 (2021). 586 29. Quesada, R. et al. The allometry of CNS size and consequences of miniaturization in 587 orb-weaving and cleptoparasitic spiders. Arthropod Struct Dev 40, 521–529 (2011). 588 30. Polilov, A. A. & Makarova, A. A. The scaling and allometry of organ size associated 589 with miniaturization in insects: A case study for Coleoptera and Hymenoptera. Sci 590 Rep 7, 43095 (2017). 591 31. Godfrey, R. K., Swartzlander, M. & Gronenberg, W. Allometric analysis of brain cell 592 number in Hymenoptera suggests ant brains diverge from general trends. Proc Biol 593 Sci 288, (2021). 594 32. Sayol, F. et al. Feeding specialization and longer generation time are associated with 595 relatively larger brains in bees. Proceedings of the Royal Society B 287, (2020). 596 33. Lösel, P. D. et al. Natural variability in bee brain size and symmetry revealed by 597 micro-CT imaging and deep learning. PLoS Comput Biol 19, e1011529 (2023). 598 34. Gowda, V. & Gronenberg, W. Brain composition and scaling in social bee species 599 differing in body size. Apidologie 50, 779–792 (2019). 600 35. Wehner, R., Fukushi, T. & Isler, K. On being small: Brain allometry in ants. Brain 601 Behav Evol 69, 220–228 (2007). 602 36. Seid, M. A., Castillo, A. & Wcislo, W. T. The allometry of brain miniaturization in ants. 603 Brain Behav Evol 77, 5–13 (2011). 604 37. Mink, J. W., Blumenschine, R. J. & Adams, D. B. Ratio of central nervous system to 605 body metabolism in vertebrates: Its constancy and functional basis. Am J Physiol 241, 606 203–212 (1981). 607 38. Kern, M. J. Metabolic rate of the insect brain in relation to body size and phylogeny. 608 Comparative Biochemistry and Physiology Part A 81, 501–506 (1985). 609 39. Chittka, L. & Niven, J. Are bigger brains better? Curr Biol 19, R995–R1008 (2009). 610 40. White, C. R., Alton, L. A., Bywater, C. L., Lombardi, E. J. & Marshall, D. J. Metabolic 611 scaling is the product of life-history optimization. Science (1979) 377, 834–839 (2022). 612 41. Harrison, J. F. Do performance-safety tradeoffs cause hypometric metabolic scaling in 613 animals? Trends Ecol Evol 32, 653–664 (2017). 614 42. Glazier, D. S. Does death drive the scaling of life? Biological Reviews 100, 586–619 615 (2025). 616 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 19 43. Reader, S. M. & Laland, K. N. Social intelligence, innovation, and enhanced brain size 617 in primates. Proceedings of the National Academy of Sciences 99, 4436–4441 (2002). 618 44. Deaner, R. O., Isler, K., Burkart, J. & van Schaik, C. Overall brain size, and not 619 encephalization quotient, best predicts cognitive ability across non-human primates. 620 Brain Behav Evol 70, 115–124 (2007). 621 45. Herculano-Houzel, S. Numbers of neurons as biological correlates of cognitive 622 capability. Curr Opin Behav Sci 16, 1–7 (2017). 623 46. Polilov, A. A. Small is beautiful: features of the smallest insects and limits to 624 miniaturization. Annu Rev Entomol 60, 103–121 (2015). 625 47. Niven, J. E. & Farris, S. M. Miniaturization of nervous systems and neurons. Current 626 Biology 22, R323–R329 (2012). 627 48. Sengupta, B., Stemmler, M., Laughlin, S. B. & Niven, J. E. Action potential energy 628 efficiency varies among neuron types in vertebrates and invertebrates. PLoS Comput 629 Biol 6, e1000840 (2010). 630 49. Attwell, D. & Laughlin, S. An energy budget for signaling in the grey matter of the 631 brain. J Cereb Blood Flow Metab 21, (2001). 632 50. Quintela-López, T., Shiina, H. & Attwell, D. Neuronal energy use and brain evolution. 633 Current Biology 32, R650–R655 (2022). 634 51. Farris, S. M. Insect societies and the social brain. Curr Opin Insect Sci 15, 1–8 (2016). 635 52. O’Donnell, S. et al. Distributed cognition and social brains: Reductions in mushroom 636 body investment accompanied the origins of sociality in wasps (Hymenoptera: 637 Vespidae). Proceedings of the Royal Society B: Biological Sciences 282, 20150791 638 (2015). 639 53. Godfrey, R. K. & Gronenberg, W. Brain evolution in social insects: Advocating for the 640 comparative approach. Journal of Comparative Physiology A 2019 205:1 205, 13–32 641 (2019). 642 54. Feinerman, O. & Traniello, J. F. A. Social complexity, diet, and brain evolution: 643 modeling the effects of colony size, worker size, brain size, and foraging behavior on 644 colony fitness in ants. Behav Ecol Sociobiol 70, 1063–1074 (2016). 645 55. Traniello, J. F., Linksvayer, T. A. & Coto, Z. N. Social complexity and brain evolution: 646 insights from ant neuroarchitecture and genomics. Curr Opin Insect Sci 53, 100962 647 (2022). 648 56. Coto, Z. N. & Traniello, J. F. A. Social brain energetics: Ergonomic efficiency, 649 neurometabolic scaling, and metabolic polyphenism in ants. Integr Comp Biol 62, 650 1471–1478 (2022). 651 57. Lihoreau, M., Latty, T. & Chittka, L. An exploration of the social brain hypothesis in 652 insects. Front Physiol 3, 442 (2012). 653 58. Traniello, J. F. A. & Avarguès-Weber, A. Individual and collective cognition in social 654 insects: What’s in a name? Behav Ecol Sociobiol 77, 1–12 (2023). 655 59. Bell-Roberts, L. et al. Larger colony sizes favoured the evolution of more worker 656 castes in ants. Nature Ecology & Evolution 2024 8:10 8, 1959–1971 (2024). 657 60. Matte, A. & LeBoeuf, A. C. Innovation in ant larval feeding facilitated queen–worker 658 divergence and social complexity. Proceedings of the National Academy of Sciences 659 122, e2413742122 (2025). 660 61. Holbrook, C. T., Barden, P. M. & Fewell, J. H. Division of labor increases with colony 661 size in the harvester ant Pogonomyrmex californicus. Behavioral Ecology 22, 960–662 966 (2011). 663 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 20 62. Anderson, C. & McShea, D. W. Individual versus social complexity, with particular 664

to ant colonies. Biol Rev Camb Philos Soc 76, 211–237 (2001). 665 63. Muratore, I. B., Fandozzi, E. M. & Traniello, J. F. A. Behavioral performance and 666 division of labor influence brain mosaicism in the leafcutter ant Atta cephalotes. 667 Journal of Comparative Physiology A 208, 325–344 (2022). 668 64. Kamhi, J. F., Gronenberg, W., Robson, S. K. A. & Traniello, J. F. A. Social complexity 669 influences brain investment and neural operation costs in ants. Proceedings of the 670 Royal Society B: Biological Sciences 283, 20161949 (2016). 671 65. Pahlke, S., Seid, M. A., Jaumann, S. & Smith, A. The loss of sociality is accompanied 672 by reduced neural investment in mushroom body volume in the sweat bee Augochlora 673 pura (Hymenoptera: Halictidae). Ann Entomol Soc Am 114, 637–642 (2020). 674 66. Smith, A. R., Seid, M. A., Jiménez, L. C. & Wcislo, W. T. Socially induced brain 675 development in a facultatively eusocial sweat bee Megalopta genalis (Halictidae). 676 Proceedings of the Royal Society B: Biological Sciences 277, 2157 (2010). 677 67. Strausfeld, N. J., Hansen, L., Li, Y., Gomez, R. S. & Ito, K. Evolution, discovery, and 678 interpretations of arthropod mushroom bodies. Learning & Memory 5, 11–37 (1998). 679 68. Riveros, A. J., Seid, M. A. & Wcislo, W. T. Evolution of brain size in class-based 680 societies of fungus-growing ants (Attini). Anim Behav 83, 1043–1049 (2012). 681 69. Jerison, H. J. Brain evolution: New light on old principles. Science (1979) 170, 1224–682 1225 (1970). 683 70. Cole, A. C. Pogonomyrmex Harvester Ants. A Study of the Genus in North America. 684 (University of Tennessee Press, Knoxville, Tennessee, 1968). 685 71. Barron, A. B. & Mourmourakis, F. The relationship between cognition and brain size 686 or neuron number. Brain Behav Evol 99, 109–122 (2024). 687 72. Graber, L. C., Johnson, R. A. & Moreau, C. S. UCE phylogenomics inform the 688 systematics and geographic range evolution of the harvester ant genus 689 Pogonomyrmex. bioRxiv 2024.11.12.623263 (2024) doi:10.1101/2024.11.12.623263. 690 73. Burger, J. R., Hou, C., A S Hall, C. & Brown, J. H. Universal rules of life: Metabolic 691 rates, biological times and the equal fitness paradigm. Ecol Lett 24, 1262–1281 692 (2021). 693 74. Brown, J. H., Burger, J. R., Hou, C. & Hall, C. A. S. The pace of life: Metabolic energy, 694 biological time, and life history. Integr Comp Biol 62, 1479–1491 (2022). 695 75. Hölldobler, B. Recruitment behavior, home range orientation and territoriality in 696 harvester ants, Pogonomyrmex. Behav Ecol Sociobiol 1, 3–44 (1976). 697 76. Whitford, W. G. Foraging in seed-harvester ants Pogonomyrmex spp. Ecology 59, 698 185–189 (1978). 699 77. Flanagan, T. P., Letendre, K., Burnside, W. R., Fricke, G. M. & Moses, M. E. 700 Quantifying the effect of colony size and food distribution on harvester ant foraging. 701 PLoS One 7, e39427 (2012). 702 78. Hölldobler, B. Recruitment behavior, home range orientation and territoriality in 703 harvester ants, Pogonomyrmex. Behav Ecol Sociobiol 1, 3–44 (1976). 704 79. Warburg, I., Whitford, W. G. & Steinberger, Y. Colony size and foraging strategies in 705 desert seed harvester ants. J Arid Environ 145, 18–23 (2017). 706 80. Weier, J. A. & Jr, D. H. F. Foraging in the seed-harvester ant genus Pogonomyrmex: 707 Are energy costs important? 291–300 (1995). 708 81. Fewell, J. H. Energetic and time costs of foraging in harvester ants, Pogonomyrmex 709 occidentalis. Behav Ecol Sociobiol 22, 401–408 (1988). 710 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 21 82. Davidson, D. W. Foraging ecology and community organization in desert seed-eating 711 ants. Ecology 58, 725–737 (1977). 712 83. Davidson, D. W. Species diversity and community organization in desert seed-eating 713 ants. Ecology 58, 711–724 (1977). 714 84. Wagner, D. & Gordon, D. M. Colony age, neighborhood density and reproductive 715 potential in harvester ants. Oecologia 1999 119:2 119, 175–182 (1999). 716 85. Waters, J. S., Holbrook, C. T., Fewell, J. H. & Harrison, J. F. Allometric Scaling of 717 Metabolism, Growth, and Activity in Whole Colonies of the Seed-Harvester Ant 718 Pogonomyrmex californicus. Am Nat 176, 501–510 (2010). 719 86. Fewell, J. H. & Harrison, J. F. Scaling of work and energy use in social insect 720 colonies. Behavioral Ecology and Sociobiology 2016 70:7 70, 1047–1061 (2016). 721 87. Ostwald, M. M. et al. Cooperation among unrelated ant queens provides persistent 722 growth and survival benefits during colony ontogeny. Sci Rep 11, 8332 (2021). 723 88. Waters, J. S., Ochs, A., Fewell, J. H. & Harrison, J. F. Differentiating causality and 724 correlation in allometric scaling: Ant colony size drives metabolic hypometry. 725 Proceedings of the Royal Society B 284, 20162582 (2017). 726 89. Duell, M. E., Klok, C. J., Roubik, D. W. & Harrison, J. F. Size-dependent scaling of 727 stingless bee flight metabolism reveals an energetic benefit to small body size. Integr 728 Comp Biol 62, 1429–1438 (2022). 729 90. Wagner, J. M. et al. Isometric spiracular scaling in scarab beetles - implications for 730 diffusive and advective oxygen transport. Elife 11, e82129 (2022). 731 91. Foerster, S. Í. A., Javoiš, J., Holm, S. & Tammaru, T. Predicting insect body masses 732 based on linear measurements: A phylogenetic case study on geometrid moths. 733 Biological Journal of the Linnean Society 141, 71–86 (2024). 734 92. Herrera, C. M. Thermal biology diversity of bee pollinators: Taxonomic, phylogenetic, 735 and plant community-level correlates. Ecol Monogr 94, e1625 (2024). 736 93. Taber, S. Welton. The World of the Harvester Ants. (Texas A & M University Press, 737 College Station, 1998). 738 94. Harrison, J. F. Approaches for testing hypotheses for the hypometric scaling of 739 aerobic metabolic rate in animals. Am J Physiol Regul Integr Comp Physiol 315, 740 R879–R894 (2018). 741 95. Harrison, J. F., Klok, C. J. & Waters, J. S. Critical PO2 is size-independent in insects: 742 Implications for the metabolic theory of ecology. Curr Opin Insect Sci 4, 54–59 (2014). 743 96. Burns, J. G., Foucaud, J. & Mery, F. Costs of memory: Lessons from ‘mini’ brains. 744 Proceedings of the Royal Society B: Biological Sciences 278, 923–929 (2011). 745 97. Shik, J. Z. & Kaspari, M. Lifespan in male ants linked to mating syndrome. Insectes 746 Soc 56, 131–134 (2009). 747 98. Turza, F., Stec, D., Fontaneto, D. & Miler, K. Life expectancy in ants explains variation 748 in helpfulness regardless of phylogenetic relatedness. Behavioral Ecology 749 https://doi.org/10.1093/beheco/arae104 (2024) doi:10.1093/beheco/arae104. 750 99. Koch, R. E. et al. Integrating mitochondrial aerobic metabolism into ecology and 751 evolution. Trends Ecol Evol 36, 321–332 (2021). 752 100. Shilovsky, G. A., Putyatina, T. S. & Markov, A. V. Evolution of longevity in tetrapods: 753 Safety is more important than metabolism level. Biochemistry (Moscow) 89, 322–340 754 (2024). 755 101. Perl, C. D. & Niven, J. E. Colony-level differences in the scaling rules governing wood 756 ant compound eye structure. Sci Rep 6, 24204 (2016). 757 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 22 102. Chong, K. L., Grahn, A., Perl, C. D. & Sumner-Rooney, L. Allometry and ecology 758 shape eye size evolution in spiders. Current Biology 34, 3178-3188.e5 (2024). 759 103. Brooke, M. D. L., Hanley, S. & Laughlin, S. B. The scaling of eye size with body mass 760 in birds. Proceedings of the Royal Society of London B: Biological Sciences 226, 405–761 412 (1999). 762 104. Huang, C. H., Zhong, M. J., Liao, W. B. & Kotrschal, A. Investigating the role of body 763 size, ecology, and behavior in anuran eye size evolution. Evol Ecol 33, 585–598 764 (2019). 765 105. Aydın, M. Ş. et al. Active shrinkage protects neurons following axonal transection. 766 iScience 26, 107715 (2023). 767 106. Oikawa, S. & Itazawa Yasuo. Relationship between summated tissue respiration and 768 body size in a marine teleost, the porgy Pagrus major. Fisheries Science 69, 687–694 769 (2003). 770 107. Oikawa, S. & Itazawa, Y. Relationship between metabolic rate in vitro and body mass 771 in a marine teleost, porgy Pagrus major. Fish Physiol Biochem 10, 177–182 (1992). 772 108. Buschiazzo, A. et al. Effect of starvation on brain glucose metabolism and 18F-2-773 fluoro-2-deoxyglucose uptake: An experimental in-vivo and ex-vivo study. EJNMMI 774 Res 8, (2018). 775 109. Moncada-Restrepo, M. & Chambers, J. W. Respirometry parameters as indicators of 776 neurotoxicity. Adv Neurotoxicol 13, 55–80 (2025). 777 110. Mackert, O. et al. Impact of metabolic stress induced by diets, aging and fasting on 778 tissue oxygen consumption. Mol Metab 64, (2022). 779 111. Underwood, E., Redell, J. B., Zhao, J., Moore, A. N. & Dash, P. K. A method for 780 assessing tissue respiration in anatomically defined brain regions. Scientific Reports 781 2020 10:1 10, 1–14 (2020). 782 112. Gordon, D. M. Chapter 4: Colony Size. in Ant Encounters: Interaction Networks and 783 Colony Behavior 75–95 (Princeton University Press, 2010). 784 113. Gordon, D. M. The rewards of restraint in the collective regulation of foraging by 785 harvester ant colonies. Nature 498, (2013). 786 114. Nielsen, M. G., Jensen, T. F. & Holm-Jensen, I. Effect of load carriage on the 787 respiratory metabolism of running worker ants of Camponotus herculeanus 788 (Formicidae). Oikos 39, 137 (1982). 789 115. Duncan, F. D. & Lighton, J. R. B. The burden within: The energy cost of load carriage 790 in the honeypot ant, Myrmecocystus. Physiol Zool 67, 190–203 (1994). 791 116. Vogt, J. T. & Appel, A. G. Standard metabolic rate of the fire ant, Solenopsis invicta 792 Buren: effects of temperature, mass, and caste. J Insect Physiol 45, 655–666 (1999). 793 117. Nielsen, M. G. & Christian, K. A. The mangrove ant, Camponotus anderseni, switches 794 to anaerobic respiration in response to elevated CO2 levels. J Insect Physiol 53, 505–795 508 (2007). 796 118. Baroni-Urbani, C. & Nielsen, M. G. Energetics and foraging behaviour of the 797 European seed harvesting ant Messor capitatus: II. Do ants optimize their harvesting? 798 Physiol Entomol 15, 449–461 (1990). 799 119. Lighton, J. R. B. Measuring Metabolic Rates: A Manual for Scientists. (Oxford 800 University Press, 2019). 801 120. Neville, K. E. et al. A novel ex vivo method for measuring whole brain metabolism in 802 model systems. J Neurosci Methods 296, 32–43 (2018). 803 121. Rittschof, C. C. & Schirmeier, S. Insect models of central nervous system energy 804 metabolism and its links to behavior. Glia 66, 1160–1175 (2018). 805 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint 23 122. R Core Team. R: A language and environment for statistical computing. Preprint at 806 https://www.r-project.org (2024). 807 123. Revell, L. J. phytools: an R package for phylogenetic comparative biology (and other 808 things). Methods Ecol Evol 3, 217–223 (2012). 809 124. Paradis, E. & Schliep, K. ape 5.0: An environment for modern phylogenetics and 810 evolutionary analyses in R. Bioinformatics 35, 526–528 (2019). 811 125. Wickham, H. Ggplot2: Elegant Graphics for Data Analysis. (Springer-Verlag New 812 York, 2016). 813 126. Yu, G. Using ggtree to visualize data on tree-like structures. Curr Protoc 814 Bioinformatics 69, e96 (2020). 815 127. Wood, S. N. Fast stable restricted maximum likelihood and marginal likelihood 816 estimation of semiparametric generalized linear models. J R Stat Soc Series B Stat 817 Methodol 73, 3–36 (2011). 818 128. Pedersen, E. & Simpson, G. MRFtools: Tools for constructing and plotting Markov 819 Random Fields in R for Graphical Data. R package version 0.0-2, commit 820 57dc21b8d46050af443d5a3c9d397e425f09914b. https://github.com/eric-821 pedersen/MRFtools (2023). 822 129. Bates, D., Mächler, M., Bolker, B. M. & Walker, S. C. Fitting linear mixed-effects 823 models using lme4. J Stat Softw 67, 1–48 (2015). 824 130. Kuznetsova, A., Brockhoff, P. B. & Christensen, R. H. B. lmerTest package: Tests in 825 linear mixed effects models. J Stat Softw 82, 1–26 (2017). 826 131. Cole, A. C. Studies of New Mexico ants. IX. Pogonomyrmex apache (Wheeler) a 827 synonym of Pogonomyrmex sancti-hyacinthi (Wheeler) (Hymenoptera: Formicidae). 828 Journal of the Tennessee Academy of Sciences 29, 266–271 (1954). 829 132. Tschinkel, W. R. Lifespan, age, size-specific mortality and dispersion of colonies of 830 the Florida harvester ant, Pogonomyrmex badius. Insectes Soc 64, 285–296 (2017). 831 133. Gordon, D. M. Nest relocation in harvester ants. Ann Entomol Soc Am 85, 44–47 832 (1992). 833 134. Creighton, W. S. Studies on the North American representatives of Ephebomyrmex 834 (Hymenoptera: Formicidae). Psyche (Camb Mass) 63, 54–66 (1956). 835 135. Creighton, W. S. Studies on Arizona ants (3) the habits of Pogonomyrmex 836 huachucanus (Wheeler) and a description of the sexual castes. Psyche (Camb Mass) 837 59, 71–81 (1952). 838 136. Heinze, J., Hölldobler, B. & Cover, S. P. Queen polymorphism in the North American 839 harvester ant, Ephebomyrmex imberbiculus. Insectes Soc 39, 267–273 (1992). 840 137. Lavigne, R. J. Bionomics and nest structure of Pogonomyrmex occidentalis 841 (Hymenoptera: Formicidae). Ann Entomol Soc Am 62, 1166–1175 (1969). 842 138. Johnson, R. A., Holbrook, C. T., Strehl, C. & Gadau, J. Population and colony 843 structure and morphometrics in the queen dimorphic harvester ant, Pogonomyrmex 844 pima. Insectes Soc 54, 77–86 (2007). 845 139. MacKay, W. P. A comparison of the energy budgets of three species of 846 Pogonomyrmex harvester ants (Hymenoptera: Formicidae). Oecologia 66, 484–494 847 (1985). 848 849 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted February 2, 2026. ; https://doi.org/10.64898/2026.01.30.702859doi: bioRxiv preprint