Monday, April 16, 2018

Resurrect Mahonia sonnei!


Mahonia sonnei was named by Abrams (1934) based on a single gathering from Truckee, California.  The syntype specimens were collected by Charles F. Sonne in August 1884 and again in April 1885 (no lectotype has been as yet designated). 

McMinn pigeon-holed M. sonnei in Berberis (McMinn 1939), as was then and still recently the custom.  We now know that Mahonia is a distinct genus well separated from Berberis, and from the related genera Alloberberis and Moranothamnus in both Asia and the New World (Yu and Chung 2017).

Poor Mahona sonnei: it has died twice.

First it disappeared for 90 years: the species was never collected again until 1973 when James B. Roof and members of the California Native Plant Society organized the founding “rare plant treasure hunt” (Roof 1974).

Mahonia sonnei was listed under the U.S. Endangered Species Act of 1973 in(Federal Register 44, p. 64246, 6 November 6 1979.  In 1993, Michael Williams placed it in synonymy in the first edition of the Jepson Manual (Williams 1993), where it has remained ever since (Williams 2012, Whittemore 1997).  Then it disappeared again - Mahonia repens was delisted – officially removed from the Endangered Species Act in 2003.  Poof!  The delisting was directly a result of a floristic treatment, not a detailed published study.

Abrams (1934) and McMinn (1939) keyed M. sonnei thusly:
Mahonia sonnei: Leaflets bright-green and shining above, teeth 12-16 on each margin, merely bristle-tipped; lower surface not gray
           Mahonia repens: Leaflets dull above; teeth small, bristle-tipped, usually 12 on a side.

In revisiting the Berberidaceae of California, I reviewed Abrams’ original description, and now that we have a wealth of imaged herbarium specimens, sought to do a quick, simple test of the number of  teeth per side of a leaflet.  Marginal teeth were counted on the 3 of the 4 syntype specimens of M. sonnei (NDG19699, US2699, DS95828), and from 70 specimens selected throughout the geographic range of M. repens (Fig. 2)

Figure 1 shows a box plot of the results: M. sonnei averages 15.9 teeth per leaflet side, while M repens averaged 10.2.  The difference is significant at p=<0 .0001="" df="68).<span" style="mso-spacerun: yes;"> 
Imaged specimens of M. repens were selected from throughout its wide range in western North America (Figure 2).

Figure 1.  Box plot of number of teeth on one side of leaflets from 3 of 4 syntypes M. sonnei and from 70 randomly chosen specimens of M. repens.  The difference in means is highly significant (T-test)

Figure 2.  Geographic distribution of M. repens, based on specimens served up on SEINet


This simple comparison suggests that M. sonnei and M. repens exhibit different leaf morphometrics.  Mahonia sonnei has never been DNA sequenced.  My inspection of specimens so far suggests that, overall, M. repens and the syntype specimens M. sonnei occupy differing leaf trait morphometric multivariate space (compare Fig. 3 and Fig. 4).  Additional work on this problem is planned: the expectation is that substantial morphometric differences between M. sonnei and M. repens can be demonstrated.  However, M. sonnei needs to be sequenced.

Figure 3 and 4.  Representative leaflets of M. sonnei (below) and M. repens (above, the leaflet chosen based on the observed mean of n=10 teeth)





Preliminary conclusion: Mahonia sonnei is a distinct taxon.  The baby was thrown out with the bathwater.


Literature Cited

Abrams, Leroy.  1934.  The Mahonias of the Pacific States.  Phytologia 1(2):89-94.
McMinn, Howard E.  1939.  An Illustrated Manual of California Shrubs.  Univ. of California Press, Berkeley, CA
Roof, J.B.  1974.  Found alive: the Truckee barberry.  Four Seasons 4(4):2-18.
Whittemore, A. 1997. Berberis. Pp. 276-286 in Flora of North America Editorial Committee (eds.), Flora of North America North of Mexico, Vol. 3. New York and Oxford.
Williams, M. 1993. Berberis. Pp. 362-364 in Hickman, J.C. (ed.), The Jepson Manual: Higher Plants of California. University of California Press, Berkeley.
Williams, M. 2012. Berberis. Pp. 446-447 in Baldwin et al. (eds.), The Jepson Manual: Higher Plants of California. University of California Press, Berkeley.
Yu, C. C., & Chung, K. F. (2017). Why Mahonia? Molecular recircumscription of Berberis sl, with the description of two new genera, Alloberberis and Moranothamnus. Taxon, 66(6), 1371-1392.

1 comment:

Alexander J. Wright said...

I've been intrigued by M. sonnei for a while. My best guess so far, however, had been that M. repens at its southwestern reaches just gets taller and shinier adaxially, making M. sonnei a parallel development to M. piperiana (as it falls out in Abrams's key). (In this way of thinking, M. pumila is totally distinct and more similar to M. dictyota.)

In the nursery trade, where much M. repens is grown by seed, it appears to very frequently introgress with M. aquifolium, the resulting plants often being just somewhat taller than M. repens with leaves grey both sides. Occasionally, however, there result tall plants with shiny leaflet uppersides and pappilate lowersides, and the marginal leaflet spines reduced in number. The leaf margin on these plants is also less undulate and the spines face toward the apex. The leaf apex varies from rounded to acute, between the parents.

These distinctions - mid-height habit, adaxially glossy and abaxially dull leaflets, just-more-acute-than-rounded leaflet apices, intermediate number of marginal spines, and range where the cordillera break down into the Pacific slope - all seem intermediate between M. repens and M. aquifolium/pinnata (the distinction as currently drawn between those latter two species being entirely artificial).

I do wonder how plastic/evolutionarily labile the leaflet margin is. There exist at least two cultivars in the genus where all marginal spines are suppressed and only the terminal spine remains - 'Marvel' from a cultivated hybrid in the Orientales, and Shnilemoon, a wild-collected Occidentales clone from Santa Cruz island!

So, to my mind, both M. piperiana and M. sonnei represent mutually independent gradations from the interior M. repens into the diverse Californian members of the complex. I'm not sure where distinctions can be drawn, since there are also independent intermediate populations outside the range of those two described species (on the east side of the Cascades in northern Oregon, for example). Why they don't seem to occur much elsewhere could be a good counterargument.

(And since I waste no opportunity to point it out, while the conclusion reached by Yu & Chung makes perfect sense to me, their key does not - the B. aquifolium complex most certainly does have dimorphic stems! One hopes that the error will not filter into future keys.)