A century of chromosome evolution research.

G. Ledyard Stebbins, in his landmark 1950 synthesis Variation and Evolution in Plants, framed polyploidy (whole-genome duplication) and dysploidy (single-chromosome gains and losses) as paths to reproductive isolation. The view dominated for decades: polyploidy was the engine of speciation in plants, full stop. The limitation, visible only later, was that the framework had no formal way to estimate rates from a phylogeny. It was a mechanism without an inference procedure.

M.J.D. White, in Animal Cytology and Evolution (1954, revised 1973), documented the astonishing diversity of animal karyotypes and advanced "stasipatric speciation", the proposal that chromosomal rearrangements could drive speciation even without geographic isolation. The idea generated fierce debate and never reached consensus. What endured was the breadth of the catalogue itself. White established that karyotype mattered beyond plants, and that animal cytogenetic diversity was real, heterogeneous, and worth explaining.

Russell Lande's 1979 paper in Evolution applied diffusion theory to chromosomal rearrangements and showed that underdominant variants, which reduce heterozygote fitness, face severe barriers to fixation in large populations. This was a direct theoretical challenge to White's stasipatric model. For a decade the field retreated. Allozymes and then DNA sequences took over the speciation conversation, and chromosome-number evolution largely disappeared from the front of evolutionary biology.

Max King's 1993 Species Evolution: The Role of Chromosome Change re-argued the case for structural change as a speciation force, and Loren Rieseberg's experimental sunflower work through the 1990s and 2000s reframed the mechanism. Chromosomal rearrangements were shown to accumulate predictably during hybrid speciation and to act as barriers to gene flow through recombination suppression rather than underdominance alone. The theoretical block Lande had raised was circumvented, and karyotype change returned as a tractable speciation mechanism.

Itay Mayrose and colleagues released ChromEvol (2010, 2014), providing the first statistically rigorous framework for inferring rates of polyploidy and dysploidy from a phylogeny. The provocative 2011 Science paper that followed, showing recently formed polyploids diversify more slowly than diploid relatives, challenged decades of received wisdom about polyploidy as a speciation engine and was re-litigated throughout the decade. Whatever the verdict, the method itself permanently changed what counted as a claim about chromosome evolution.

Will Freyman and Sebastian Höhna's ChromoSSE (2018) extended the probabilistic program by letting chromosome number change interact with speciation and extinction rates directly. Rosana Zenil-Ferguson's BiChroM built in parallel, coupling chromosome evolution to trait evolution. These joint models treat the karyotype as dynamic and causally important, rather than as a passive marker. Their assumptions are under active scrutiny, as all state-dependent models are, but they defined the methodological edge that the current literature is still arguing about.

The probabilistic methods of Mayrose and Freyman & Höhna were built for plants, where curated chromosome-count datasets already existed. Animals lagged badly. The emergence of course-based undergraduate research experiences as a data collection pipeline, together with open cytogenetic databases for beetles, flies, amphibians, and mammals, has changed that. Animal comparative work is now possible at scales that were out of reach a decade ago, and the results are starting to show that stasis in groups like birds was an artifact of undersampling, not biology.

Language models have made it possible to extract structured cytogenetic claims from the primary literature at a rate no graduate student can match. TraitTrawler and related agents are not replacing the careful curation the field requires, but they are reshaping what the curation job looks like: the humans now validate and adjudicate rather than read every paper from cover to cover. The bet is that this turns the historical bottleneck, which was always data extraction and never research questions, into something a lab of reasonable size can actually clear.

Three threads are unresolved. Whether state-dependent diversification methods can be trusted without hidden-state correction is still contested (Beaulieu & O'Meara 2016; Caetano et al. 2018). Whether our rate models capture the biology well enough to trust the inferences, the model adequacy problem, is only starting to be tested. And the animal literature, despite the CUREs-era data, has yet to produce a probabilistic synthesis of the scope Mayrose and Zenil-Ferguson built for plants. The next decade's work is almost certainly on integrating mechanistic cytogenetics (centromere biology, recombination landscape) with macroevolutionary inference.