Black ash restoration is a sourcing problem: 14% of seed variation is regional, 43% within a stand

The two variance figures, as reported, do not sum to the whole. Within-population variation averaged 43% and between-population variation 14% ^[5][6], which leaves 43% of the measured seed-trait variation attributed to neither level in the account of the work ^[7]. The residual matters less than the ratio between the two numbers that are given: trees growing in the same population differ about three times as much as one population differs from another ^[8]. Regional divergence is real, and the team attributes it to climate ^[3]. It is also the smaller of the two signals.

That creates an awkward operating rule for anyone drawing on these collections. Matching a planting to its source climate is necessary, because seed traits govern how seedlings develop and how well they survive and grow ^[3], and the team ran climate modelling specifically to predict how seed from one climate fares in another ^[18]. But most of the diversity a restored stand would need sits inside each provenance, which means the worth of a seed lot depends on how many maternal trees it represents. That number is not in the reported split. The imaging covered more than 700 lots and about 35,000 seeds ^[10], which is no more than roughly 50 seeds per lot ^[11], and a lot is not a mother tree.

The choice of X-ray imaging is the quiet load-bearing decision here: it reads the inside of a seed without destroying it ^[10]. For populations that no longer stand anywhere on the landscape ^[15], destructive phenotyping spends the only copy.

The three gardens do the causal work. Traits that stay apart when Pennsylvania, Syracuse and Quebec City material is grown side by side are genetic; traits that converge were environmental ^[13][14]. They were planted from 22,000 germinated seedlings ^[12] and may eventually produce restoration seed themselves ^[17]. That is where sourcing policy gets set without being announced, because whatever provenance mix went into those gardens becomes the provenance mix of everything grown from them.

Jill Hamilton's ark framing is precise in one respect worth taking literally: many of the populations in these collections are trees that are no longer on the landscape, and the germplasm is wanted for borer-resistance breeding as well as restoration ^[15]. So the binding constraint is not seed volume. It is whether the provenance a given wetland needs was collected before the beetle arrived ^[1], and whether resistance selection leaves that provenance in the pool. Black ash pays out locally: it regulates hydrology in the wetland where it stands, and it supplies basketry material to Wabanaki, Iroquois and Great Lakes communities near particular stands ^[16]. A species that persists in aggregate but not in those places has not been restored.