Royal Botanic Gardens Kew, Richmond, Surrey, UK.
J Exp Bot. 2021 Feb 2;72(2):654-681. doi: 10.1093/jxb/eraa467.
Plastid sequences have long dominated phylogeny reconstruction at all time depths, predicated on a usually untested assumption that they accurately represent the evolutionary histories of phenotypically circumscribed species. We combined detailed in situ morphometrics (124 plants) and whole-plastome sequencing through genome skimming (71 plants) in order to better understand species-level diversity and speciation in arguably the most challenging monophyletic group within the taxonomically controversial, pseudo-copulatory bee orchid genus Ophrys. Using trees and ordinations, we interpreted the data at four nested demographic levels-macrospecies, mesospecies, microspecies, and local population-seeking the optimal level for bona fide species. Neither morphological nor molecular discontinuities are evident at any level below macrospecies, the observed overlap among taxa suggesting that both mesospecies and microspecies reflect arbitrary division of a continuum of variation. Plastomes represent geographic location more strongly than taxonomic assignment and correlate poorly with morphology, suggesting widespread plastid capture and possibly post-glacial expansion from multiple southern refugia. As they are rarely directly involved in the speciation process, plastomes depend on extinction of intermediate lineages to provide phylogenetic signal and so cannot adequately document evolutionary radiations. The popular 'ethological' evolutionary model recognizes as numerous 'ecological species' (microspecies) lineages perceived as actively diverging as a result of density-dependent selection on very few features that immediately dictate extreme pollinator specificity. However, it is assumed rather than demonstrated that the many microspecies are genuinely diverging. We conversely envisage a complex four-dimensional reticulate network of lineages, generated locally and transiently through a wide spectrum of mechanisms, but each unlikely to maintain an independent evolutionary trajectory long enough to genuinely speciate by escaping ongoing gene flow. The frequent but localized microevolution that characterizes the Ophrys sphegodes complex is often convergent and rarely leads to macroevolution. Choosing between the contrasting 'discontinuity' and 'ethology' models will require next-generation sequencing of nuclear genomes plus ordination of corresponding morphometric matrices, seeking the crucial distinction between retained ancestral polymorphism-consistent with lineage divergence-and polymorphisms reflecting gene flow through 'hybridization'-more consistent with lineage convergence.
质体序列长期以来一直主导着所有时间深度的系统发育重建，这是基于一个通常未经测试的假设，即它们准确地代表了表型上界定的物种的进化历史。我们结合了详细的原位形态计量学（124 株植物）和通过基因组掠过的全质体测序（71 株植物），以便更好地理解在分类学上有争议的、拟交配的蜜蜂兰属 Ophrys 中最具挑战性的单系群内的物种水平多样性和物种形成。通过使用树和排序，我们在四个嵌套的种群动态水平上解释了数据——宏观种、中观种、微观种和局部种群——寻找真正的物种的最佳水平。在任何低于宏观种的水平上都没有明显的形态或分子不连续性，观察到的分类群之间的重叠表明，中观种和微观种都反映了连续变化的任意划分。质体比分类分配更强烈地代表地理位置，与形态相关性差，表明广泛的质体捕获和可能来自多个南部避难所的后冰河时代扩张。由于它们很少直接参与物种形成过程，质体依赖于中间谱系的灭绝来提供系统发育信号，因此不能充分记录进化辐射。流行的“行为”进化模型认识到许多“生态物种”（微观种）谱系，由于对少数直接决定极端传粉者特异性的特征的密度依赖选择而被认为是积极分化的。然而，人们假设而不是证明许多微观种确实在分化。相反，我们设想了一个复杂的、具有四个维度的、由谱系组成的网状网络，这些谱系是通过广泛的机制在当地和暂时产生的，但每个谱系都不太可能通过逃避持续的基因流来维持足够长的独立进化轨迹，从而真正形成物种。Ophrys sphegodes 复合体所具有的频繁但局部的微观进化通常是趋同的，很少导致宏观进化。在对比鲜明的“不连续性”和“行为学”模型之间做出选择将需要对核基因组进行下一代测序，并对相应的形态计量矩阵进行排序，以寻求保留祖先多态性与谱系分化一致的关键区别，以及反映通过“杂交”产生基因流的多态性，与谱系趋同更为一致。
