Finarelli John A, Flynn John J
Committee on Evolutionary Biology, The University of Chicago, 1025 E. 57th St., Chicago, Illinois 60637, USA.
Evolution. 2007 Jul;61(7):1758-72. doi: 10.1111/j.1558-5646.2007.00131.x.
A weighted-average model, which reliably estimates endocranial volume from three external measurements of the neurocranium of extant taxa in the mammalian order Carnivora, was tested for its applicability to fossil taxa by comparing model-estimated endocranial volumes to known endocast volumes. The model accurately reproduces endocast volumes for a wide array of fossil taxa across the crown radiation of the Carnivora, three stem carnivoramorphan taxa, and Pleistocene fossils of two extant species. Applying this model to fossil taxa without known endocast volumes expanded the sample of fossil taxa with estimated brain volumes in the carnivoran suborder Caniformia from 11 to 60 taxa. This then allowed a comprehensive assessment of the evolution of relative brain size across this clade. An allometry of brain volume to body mass was calculated on phylogenetically independent contrasts for the set of extant taxa, and from this, log-transformed encephalization quotients (logEQs) were calculated for all taxa, extant, and fossil. A series of Mann-Whitney tests demonstrated that the distributions of logEQs for taxa early in caniform evolutionary history possessed significantly lower median logEQs than extant taxa. Median logEQ showed a pronounced shift around the Miocene-Pliocene transition. Support tests, based on likelihood ratios, demonstrated that the variances of these distributions also were significantly lower than among modern taxa, but logEQ variance increased gradually through the history of the clade, not abruptly. Reconstructions of ancestral logEQs using weighted squared-change parsimony demonstrate that increased encephalization is observed across all major caniform clades (with the possible exception of skunks) and that these increases were achieved in parallel, although an "ancestor-descendant differencing" method could not rule out drift as a hypothesis. Peculiarities in the estimated logEQs for the extinct caniform family Amphicyonidae were also investigated; these unusual patterns are likely due to a unique allometry in scaling brain to body size in this single clade.
一种加权平均模型能够根据食肉目现存类群颅骨的三项外部测量可靠地估算颅内容积,通过将模型估算的颅内容积与已知的脑腔模型容积进行比较,测试了该模型对化石类群的适用性。该模型能准确再现食肉目冠群辐射范围内众多化石类群、三个食肉形亚目干群类群以及两个现存物种的更新世化石的脑腔模型容积。将此模型应用于脑腔模型容积未知的化石类群,使犬形亚目化石类群中估算脑容量的样本从11个类群增加到60个类群。这进而能够全面评估该类群相对脑大小的演化。基于系统发育独立对比,计算了现存类群的脑容量与体重的异速生长关系,并据此为所有现存和化石类群计算了对数转换后的脑化商(logEQs)。一系列曼-惠特尼检验表明,犬形类进化历史早期类群的logEQs分布中位数显著低于现存类群。中位数logEQ在中新世-上新世过渡时期出现明显变化。基于似然比的支持检验表明,这些分布的方差也显著低于现代类群,但logEQ方差在该类群历史中逐渐增加,而非突然增加。使用加权平方变化简约法重建祖先logEQs表明,在所有主要犬形类分支中(可能臭鼬除外)都观察到脑化增加,且这些增加是并行实现的,尽管“祖先-后代差异”方法不能排除漂变作为一种假设。还研究了已灭绝的犬形类两栖犬科估计的logEQs的特殊性;这些异常模式可能是由于该单一分支中脑与身体大小缩放的独特异速生长关系所致。