Plant Species of the Borrego Desert: Nyctaginaceae: Allionia incarnata, trailing four o'clock
Allionia incarnata was first described in 1759, with var. denudata first described (as a separate species) in 1844, and var. villosa first described in 1909. Munz 1974, and the 1993 Jepson Manual, did not recognize the varieties of Allionia incarnata.
Turner published a "Revisionary Study of the Genus Allionia" in Phytologia in 1994, in which he restored those varieties, writing:
Allionia incarnata appears to have three well-marked but intergrading morphogeographical infraspecific taxa in North America:
- var. incarnata, a widespread weedy taxon primarily centered in North America but extending into South America;
- var. nudata largely confined to the Sonoran desert regions of Mexico and closely adjacent USA (southern California, southern Nevada, and southwestern Arizona; and
- var. villosa, largely confined to the intermontane region of Nevada and Utah southwards to northern Mexico.
The 2003 Flora of North American treatment (FNA) restored var. villosa, but synonymized var. nudata with var. incarnata. The 2012 Jepson Manual revision followed the FNA treatment.
I had seen the draft 2012 Jepson Manual revision giving the varieties online in 2010, and was a bit skeptical since I had never noticed much difference in the Borrego Desert Allionia flowers. However, on 19 April 2010, I saw an Allionia incarnata with a WHOPPING flower at Scissors Crossing I could hardly believe it was the same allionia I'd seen elsewhere in the Borrego Desert! I measured the heads to be 16 to 24 mm in corolla width, compared to heads of 6 to 10 mm I had measured elsewhere.
So I was suddenly convinced these varieties made sense.
Since then, every so often I have been measuring the range in size of the flowers in populations I see, and occasionally photographing the corolla width with a ruler next to the corolla.
A lot of my data is still buried in my notebooks, but Fig. 1 shows the range data I have digitized so far.
Fig. 1. The range of corolla widths from 12 separate populations measured in the field. Some populations plot on top of each other, so only 9 data points appear in the plot. For example, there are two separate populations that each had heads of 15 to 19 mm in width.
From these data, it appears there are clearly two separate varieties, one with heads 7 to 12 mm in width, and another with heads 13 to 24 mm in width. This is fairly consistent with the "mostly 4 to 8 (10) mm" and the "10 to 20 mm" given in Turner to separate the two varieties. However, it is also possible that the two points at the low end of the var. villosa population (13 to 14 mm, and 13 to 17 mm) could be intermediate populations. More data are needed to confirm or deny that speculation.
Note that my measurements are of corolla width of live flowers in the field, whereas the measurements in Turner are on pressed, dried specimens. Dried specimens usually exhibit some corolla shrinkage, but specimens of heads can sometimes get wider when pressed. I'll press some specimens that I've measured in the field to see what the net result is.
Oddly, though, the FNA treatment used a different break in the varieties, of "5 to 10 (12) mm" and "20 to 25 mm". This is especially surprising since the only reference cited in the FNA treatment is Turner!
Turner examined approximately 1,000 herbarium specimens, all from institutions outside of California except for UC.
Since there is no discussion in the FNA treatment about why that treatment uses such dramatically larger values for the corolla width of var. villosa, and because my data agree with the widths given by Turner, it seems unlikely that the numbers in the FNA treatment, and the Jepson Manual treatment, are correct.
For var. villosa, Turner wrote:
This taxon is weakly differentiated from var. incarnata and is largely distinguished from the latter by its generally more robust habit, longer pedicels, and larger anthers with elongate anthers. ... Still, had the taxon not already been provided with a name, I would have not been hesitant to provide one since the combination of characters which mark it seem largely confined to Arizona and closely adjacent regions. Nevertheless, intermediates between var. villosa and var. nudata occur in regions of overlap, or near overlap.
I will continue to measure population corolla widths, to see if I can find any intermediate populations, or confirm that the two lowest ranges for var. villosa are such intermediate populations.
In addition, in future updates of this page, I'll look at the "longer pedicels and elongate anthers" of var. villosa, to see how well they differentiate the varieties.