Subsequent to my initial ID efforts, I checked the 1993 'Annotated Catalogue of the Plants of the Cape Region' by Lee Lenz, and was quite puzzled when I found that neither C. californica nor C. cordifolia was listed...only C. arizonica and C. pectinata. Both the latter seemed incompatible candidates for the plant I'd found here. Among other things, C. pectinata (line drawing here) has much longer, narrower, revolute leaves than the plant in my post here; while the C. arizonica I had seen in Anza-Borrego had a totally different habit. Photos of the latter plant are posted here. The Anza-Borrego plant was a small, rounded, densely and divaricately multi-branched woody-twiggy shrub with much smaller elliptic leaves, and sessile inflorescences in the leaf axils, with smaller flowers. In contrast, the Baja plant here was more of an open perennial herb, branched from a barely suffrutescent base, with a few well-separated, virgate stems, having much larger and longer ovate-lanceolate leaves on longer petioles, which were widely spaced by long open internodes, and had long inflorescence branches located more or less distally on the shoots, with much larger flowers. It was difficult for me to conceive that these could be the same species.

After that, a long time passed before I obtained the 1983 revision of Carlowrightia by T. F. Daniel. This excellent scholarly monograph is where C. californica, C. cordifolia and others were synonymized with C. arizonica. Therein, the author explained the taxonomic difficulties he encountered in attempting to circumscribe the morphologically variable complex of entities, extending from southeastern California to Costa Rica, which he settled on treating via the 'conservative' concept of a single umbrella species, A. arizonica, consisting of 8 ''forms''. He acknowledged that his treatment was somewhat radical in that these forms, in their natural populations, appeared to be reasonably well-isolated reproductively, and that his notion of C. arizonica was not to be taken in the sense of the ''biological species concept'' of Grant (1957), but rather in the sense of the 'taxonomic species concept' of Davis & Heywood (1973). (See end of this note for bibliographic ref's.). The principal issue which led to this lumping of previously named taxa into C. arizonica appears to have been the difficulty of finding and/or formulating clear criteria that would consistently separate members of the complex, due to overlap of character states with wide variation. Apparently, Daniel did not find the variation to be sufficiently correlated to geographical distributions to merit naming subspecies, citing wide variation found to occur on both a local population level as well as on a wider geographic level.

As best I can discern, the plant in my post here would correspond to C. arizonica ''Form 4'' (see pg 81 of Daniel, 1983), which incorporated elements of the previously recognized species C. californica and C. cordifolia and, together with ''Form 5'' occurs in the Cabo San Lucas region of far southern Baja California Sur. But it's hard to know for sure, because Form 4 and Form 5 are distinguished by trichome characters not clearly discernible in my photos. Daniel also published a 1997 treatment of Acanthaceae of Baja California (available here). In that work, 5 forms of C. arizonica are treated, and again they are mainly distinguished by differences in vesture (= trichomes or ''hairs'') and geographic distribution. For that reason, I can't tell for sure which of Forms B, C, or D my plant would correspond to, although the roughly similar looking plant from the same general area in the Sanders et al voucher here is assigned to Form B. Another image from the Cabo San Lucas region showing large flowers can be seen here.

I admire the high quality of Daniel's research and scholarship, and appreciate the difficulties of addressing the morphological variation he cites in his circumscription of C. arizonica. However, I still find it very difficult to refer to the radically different plants I've seen from the Anza-Borrego Desert in southern California, USA, and the distant southern end of Baja California as the same taxon. I am very curious as to the degree and relative abundance of variants and gradations between those two extremes...on both local scales, and in terms of more gradual clines at various stations between Anza-Borrego and the Cabo San Lucas area. The discussion of C. californica here indicates that overgrazed plants were browsed down to twiggy mats whereas plants growing in the shelter of cacti produced long, loosely-branched open growth. Environmental factors like that might help explain part of the gap between plants of radically different habit. But even with impacts of browsing, different substrates, different site aspects, climate, and similar factors...it still seems a large gap to bridge between the Anza-Borrego plants and the Cabo area plants.

Daniel emphasized that his treatment was not meant as a final definitive circumscription of the elements he collected together under C. arizonica, and he hoped future work would be able to clarify boundaries of possibly distinct taxa within that grouping. While he wanted to avoid describing weakly defined taxa that might end up generating a number of superfluous synonyms for future workers, it is also desirable to avoid lumping morphologically distinct and genetically fixed entities under the same name. A number of new populations of C. arizonica have been discovered in the Anza-Borrego area in recent years. Perhaps a careful survey and study of variation vs. cohesion of characters in those populations can help further illuminate the situation?

References:

Grant, V. (1957). The plant species in theory and practice. In: Mayr, E. The Species Problem, pp. 39-80. Amer. Assoc. Adv. Sci., Washington.

Davis, P. H. & V. H. Heywood. 1973. Principles of Angiosperm Taxonomy. Krieger, Huntington, New York. xx + 558 pp.