Geographic Distribution of Arctostaphylos glandulosa subspecies in SnGb from vouchers
Keeley, Vasey and Parker 2007 (Madrono, 54:42) did a monumental study of Arctostaphylos glandulosa subspecies, measuring 52 (!!) characters for each and every one of over 1,800 (!!) voucher specimens. They did a thorough analysis of their 69 populations for which they had at least 5 individuals per population, which gave a total of 1,342 plants.
In their analysis, each population was considered as a whole, and ultimately given a single determination, even though some populations had a mixture of plants that might individually be determined as different subspecies. This makes considerable sense, since it correctly recognizes that a given interbreeding population has a range of variation. However, since most vouchers are of a single plant in a population, determinations from those individual vouchers might be at variance with the Keeley, Vasey and Parker population determination. Of course, presumably most vouchers would agree with the Keeley, Vasey and Parker population determination.
A. glandulosa is particularly difficult to partition its plants into subspecies, since populations are often an intermixture of plants with different characteristics. Keeley, Vasey and Parker point out that burl-forming species like A. glandulosa, which live through fires, have increased morphological variation compared to species that must germinate from seed after each fire, since their population isn't "reset from seeds" after each fire. One would there expect that A. glandulosa populations would vary much more than, for example, A. glauca populations.
At the end of their abstract they state:
Although all of the A. glandulosa taxa described here are known from allopatric populations, intergradations of these closely related taxa occur and thus some populations reflect a mixture of traits and can not be assigned a unique name of practical value.
In particular, they comment that one of their populations, from Crystal Lake, contains a mixture of plants with long hairs and plants without long hairs, which would each key to different subspecies.
In such cases, it is probably best just to determine the vouchers to the species, rather than try to shoehorn them into subspecies that don't actually fit them.
The four subspecies present in SnGb and their key characteristics are given in Table 1.
Table 1. The SnGb A. glandulosa subspecies and their key characteristics (with comparison to A. parryana tumescens)
| Subspecies | Branchlet hairs | Leaf hairs | Leaf color | Lower infl bracts | Fruit | Stones | Ovary |
|---|---|---|---|---|---|---|---|
| glandulosa | stalked glandular | variable | green to gray-green | mostly foliaceous, stalked glandular | slightly depressed spheric | distinct | densely white hairy and usually glandular hairy |
| mollis | non-glandular short with at least some non-glandular long hairs | glabrous | bright lustrous green | occasionally leaf-like, densely white non-glandular hairy | slightly depressed spheric | distinct | densely white hairy |
| cushingiana | non-glandular short | glabrous to pubescent | green or gray-green | leaf-like or not, densely white non-glandular hairy | slightly depressed spheric | usually distinct | densely white hairy |
| gabrielensis | non-glandular short | glabrous or sparsely puberulent | bright lustrous green | reduced or absent, densely white non-glandular hairy | slightly depressed spheric | 32-35% with single apiculate stone, mean # segments 2.0-2.4 | densely white hairy |
| A. parryana tumescens | non-glandular short | glabrous or sparsely nonglandular-hairy | bright lustrous green | reduced, glabrous (but they are actually hairy at SnGb) | spheric | 97% with single apiculate stone, mean # segments 1.1 | glabrous |
In Table 1, the characteristics used in the keys to separate the subspecies are given in bold, as well as anything else unique to one subspecies. The branchlet hairs uniquely define ssp. glandulosa (stalked glandular) and mollis (with at least some long non-glandular hairs). Both ssp. cushingiana and gabrielensis have non-glandular short hairs on their branchlets, and are separated by the leaf hairs (none for gabrielensis and none to pubescent for cushingiana) and by the color of the leaves (shiny bright green for gabrielensis and green or gray-green, not shiny, for cushingiana).
There is a discrepancy between the Keeley, Vasey and Parker paper and their Flora of North America treatment on the leaf hairs for cushingiana. The key in their paper says the leaves are slightly pubescent, whereas the FNA treatment says the leaves are glabrous or sparsely puberulent. Instead, the FNA treatment separates these two subspecies on their stones: mostly connate into single or weakly connate sphere for gabrielensis, and usually distinct for cushingiana.
As discussed above, plants that have different combinations of characteristics can not be determined to subspecies, such as plants with bright lustrous green pubescent leaves with non-glandular short branchlet hairs, or plants with equal numbers of stalked glandular hairs and non-glandular long hairs on their branchlets.
The FNA treatment says Subspecies cushingiana is widespread in mixed populations with subsp. glandulosa from Humboldt County in the North Coast Range south to the Peninsular Range, and extending into northern Baja California.
Fig. 1 maps the Keeley, Vasey and Parker populations in SnGb and gives their determinations:
Fig. 1. Map of the Keeley, Vasey and Parker SnGb populations with their determinations. The gabrielensis populations have blue circles surrounding red dots. The other red dots are labeled with their subspecies determinations.
Orlando Mistretta reviewed all the RSA vouchers in 2008, soon after the Keeley, Vasey and Parker paper came out, and redetermined each specimen. The following plots show his determinations made from individual specimens for each subspecies.
Fig. 2. Map of the ssp. gabrielensis RSA voucher locations, using the clustering algorithm at the Consortium of California Herbaria, where the number within each circle represents the number of vouchers in or near that location. The location with 22 vouchers is one of the Keeley locations where many plants were vouchered, and hence nothing is implied about whether ssp. gabrielensis actually has more plants there.
The distribution from individual vouchers of ssp. gabrielensis in Fig. 2 looks very similar to that in Fig. 1 from the Keeley, Vasey and Parker locations.
Fig. 3. Map of the ssp. mollis RSA voucher locations.
The distribution shown in Fig. 3 from individual vouchers of ssp. mollis shows a much more extensive population than from the Keeley, Vasey and Parker locations. This is because Keeley voucher locations intentionally targeted locations where collections were likely to be most informative, such as areas where Keeley knew there was variation not clearly accounted for by past taxonomic treatments; areas occupied by previously-described subspecies; and regions where there had not been much prior collecting.
Fig. 4. Map of the ssp. cushingiana RSA voucher locations.
Fig. 5. Map of the ssp. glandulosa RSA voucher locations.
Voucher data provided by the participants of the Consortium of California Herbaria (ucjeps.berkeley.edu/consortium/). retrieved on 3 October 2013.