牛科(哺乳纲:偶蹄目)动物与食物有关的适应形态模式

利用逐步分辨分析方法（Stepwise discriminant analysis，SCDA）检测了广义牛科动物的颅齿部结构，这些结构特征可以作为采食行为生态适应特征。在本研究中，测量了72种广义牛科动物的28个颅齿部结构。逐步分辨分析方法得出了6种采食方式适应类型：一般粗食者、新鲜禾草粗食者、开阔生境混合型采食者、精食者、郁闭生境混合型采食者、食果者。用103个标本检测了分辨指标的预测能力，所用标本为缺损标本，大多数缺少一项或多项结构。从这些标本获得的分辨函数的平均预测能力为94％，比用72种广义牛科动物标本建立的分辨函数的平均预测能力（98％）低一些。从一个颅齿部结构小样本建立的分辨函数可以用于考古发掘物中不完整标本的研究。这些指标与用颅下结构测量建立的运动能力和生境选择的指标相结合，可以推断古牛科动物的个体生态学以及古环境重建。

Characterizing adaptive morphological patterns related to diet in Bovidae (Mammalia: Artiodactyla)

Morphological patterns in the craniodental structure of extant bovids indicative of ecological adaptations for feeding behavior were identified by stepwise discriminant analysis (SCDA). The analyses were carried out using 28 craniodental measurements from 72 extant bovid species. The discriminant functions obtained allow the characterization of six different types of dietary adaptations: general grazers, fresh-grass grazers, mixed-feeders from open habitat, browsers,mixed-feeders from closed habitat, and frugivores. The predictive ability of the algorithms was tested with 1 063 specimens, most of them lacking one or more measurements. The mean predictive ability of the discriminant functions over these specimens (94 % ) is only slightly lower than over the species used to obtain them (98 % ). Given that the algorithms involve small sets of craniodental measurements, they can also be applied to non-complete specimens from archaeological assemblages. These algorithms, combined with those obtained using postcranial measurements for estimating adaptations related to locomotor performance and habitat choice, are useful for deriving inferences on the autecology of ancient bovids and may be used in paleoenvironmental reconstruction Acta Zoologica Sinica 52 (6): 988- 1008, 2006].