Implications of lack of morphological information in fossil leaves related to Lauraceae: a statistical evaluation

Fossil leaves serve as crucial repositories of palaeobiological information on angiosperms, with the Lauraceae family standing out as a highly represented group across geological time. The recognition of this diversity has been based mainly on comparing its leaf architecture and cuticle anatomical characteristics; however, the latter is sometimes not preserved, so leaf architecture is left as the only source of information available for identification in Lauraceae. This issue has produced a taxonomic mosaic in the family's fossil record, with taxa exhibiting well-established Lauraceae affinities and others with uncertain affinities. Nevertheless, the characters can be incompletely conserved (unknown characteristics), representing an obstacle in the identification process, which is one of the main challenges when making taxonomic assignments. Considering this problem, this study statistically analyses specific cases (it is not a taxonomic review of the family) and examines the taxonomic implications of this lack of information. The analysis used two statistical approaches: one accounting for the unknown information and another employing character imputation. A large dataset was utilized to ensure meaningful results, including extant Lauraceae taxa and fossils assigned to or related to the family. The results highlight the significant impact of unknown data on morphologic similarities and taxonomic affinities, particularly within Lauraceae, revealing challenges related to similar morphologies between extant and fossil groups and patterns found in other angiosperm groups. This study pioneered the demonstration and testing of this influential factor in taxonomic decisions, emphasizing the need for careful consideration in identifying fossil material.

化石叶片是被子植物古生物学信息的重要载体，而樟科（Lauraceae）是地质历史中化石记录占比极高的类群之一。此前对该类群多样性的识别主要依赖于对比其叶部构型与角质层解剖特征；但角质层有时难以保存，因此叶部构型成为樟科化石物种鉴定的唯一可用信息来源。这一问题导致该科的化石记录呈现分类学马赛克状分布：部分类群的樟科亲缘关系已得到明确证实，而另一部分类群的亲缘关系则尚不明确。尽管如此，分类性状往往保存不全（即特征未知），这成为物种鉴定过程中的一大障碍，也是分类学归属判定时面临的主要挑战之一。针对这一问题，本研究对具体案例开展统计学分析（本研究并非针对樟科的分类学综述），并探讨了信息缺失对分类学解读的影响。本次分析采用了两种统计学方法：一种是考虑未知信息的分析方案，另一种则采用了性状插补技术。为确保分析结果具备统计学意义，本研究使用了大规模数据集，其中涵盖现生樟科类群以及被归入或疑似归入该科的化石类群。研究结果凸显了未知数据对形态相似度与分类亲缘关系判定的显著影响，尤其在樟科类群中表现突出；该研究同时揭示了现生与化石类群间形态相似性所带来的鉴定难题，以及其他被子植物类群中存在的类似模式。本研究首次在分类学判定中对这一关键影响因素进行了验证与演示，并强调了在化石材料鉴定过程中需谨慎考量上述问题。