The direction of the pattern is where the mechanism sits. In a ZW system males are ZZ and females ZW; in an XY system females are XX and males XY [17], which means two of every three X chromosomes in a population are in a female body [4] while a Z spends most of its copies in males. An ancestral chromosome already loaded with genes preferentially expressed in the testis is better placed as a Z than as an X, and that is what the gecko comparison found [15]. Zhang Guojie's team reports the same association across 39 vertebrate species, strongest in amniotes with internal fertilisation, and reads sexual selection on sperm as the steering force [16].
That is a correlation between ancestral gene content and eventual outcome, not a demonstration that gene content causes the outcome. What makes it more than a pattern-match is that it points forward: score the testis-bias of the autosomes in a species that still uses temperature, and the rule says which of them should turn up as a Z if that lineage ever fixes a genetic switch. Geckos have crossed between environmental and genetic sex determination at least 23 times [6], so there are cases to check.
The other headline finding is thinner. Sex chromosome ages, estimated from sequence divergence, run from under 6 million years to more than 54 million [11], and the clustering claim rests on five of 11 dated species falling into a 7-to-12-million-year window around 9.6 million years ago [12], which is 45 percent of a sample of 11 [3]. Five points near a global cooling event [13] is consistent with climate pushing lineages off temperature-dependence [14], and it is also consistent with those lineages sharing an ecology, or with a shared bias in divergence-based dating. The thing this does not tell you is which.
The arithmetic in the summary also needs reconciling. The stated composition of 3 temperature-dependent, 6 XY and 11 ZW species [3] sums to 20, two short of the 22 species said to be analysed [1], and the release elsewhere describes 22 distinct sex chromosome systems in 20 species with genetic sex determination [8] against its own tally of 17 such species [2]. The paper in Science may resolve this in its species table [1]; the public account does not, and the denominator matters for the 70 percent coverage claim [4] as much as for the dating.
One result quietly raises the cost of follow-up work. Even sex chromosomes descended from the same ancestral chromosome differ in which regions became sex-differentiated, and the genes known to trigger sex determination in other vertebrates are usually missing from those regions [10]. Whatever the switches are in geckos, they will not be found by screening a candidate list borrowed from mammals.
My read, with its condition attached: the testis-gene rule is the durable part of this paper, because it replicates outside the study group and predicts a direction rather than a fit [15][16]. The Miocene birth pulse is the part I would not build on until someone dates those lineages a second way.
