Plants of Southern California: Chrysothamnus nauseosus subspecies
Table of Contents
III. Botanical Properties
IV. Geographic Distribution
We have surveyed the Angeles Crest Highway (SR2), the Angeles Forest Highway (AFH), and a handful of other roads in the eastern San Gabriel Mountains, in order to determine the distribution and properties of the four subspecies of Chrysothamnus nauseosus present here. Surprisingly, we find that two subspecies, consimilis and mohavensis, are not present in their pure forms. Instead, they form an intergrading complex whose properties are nearly homogenous throughout the area except for a general trend for longer involucres from southeast to northwest.
We find a strong dependence of bloom time on elevation. On the desert side of the mountains, all subspecies begin their bloom first at the highest elevation, with bloom progressively delayed with lower elevation. On the coastal side of the mountains, the consimilis / mohavensis complex also begin blooming at the lowest elevation, with its latest bloom at intermediate elevations. At any given elevation, each subspecies has a distinctive bloom period.
The other botanical properties of the four subspecies are largely as expected from the floras, although the corolla lengths are somewhat different than that given in the floras. It also took us quite some time to understand how the seemingly-glabrous stems of some subspecies were called tomentose or felted in the floras.
We describe the evolution of our understanding of these subspecies, and how the floras describe them, as an example of how hard it is sometimes to arrive at a correct species identification and understanding of some plant taxa.
Rabbitbrush has always been one of our favorite species in the San Gabriel Mountains for a number of reasons:
- The plants are beautiful; the roads along the desert slopes are lined with their showy yellow blooms in September and October, providing the best flower display of the fall season.
- The plants are easy to recognize, at least to the species level, being prominent shrubs, with their discoid heads with exactly five flowers per head and their phyllaries in exactly five vertical ranks.
- Finally, the plants are simply amazing, that they can bloom so well a full six months after significant rainfall in areas that receive little rainfall in the first place. How they manage to do that is no small feat, one that we constantly admire.
In 2001, as we were embarking on our quest to learn the plants of Southern California, we decided to study all the rabbitbrush plants along the Angeles Crest Highway (SR2) at its northeastern end, to see if we could find and recognize the different subspecies present in Southern California. We are quite interested in plant distributions, and so also decided to produce a map showing the range of the different subspecies.
The subspecies present in Southern California, along with their ranges in Southern California, as given in Munz (1974), are:
Subspecies Habitat Elevation Range (feet) Geographic Range bernardinus dry benches, Montane Coniferous Forest 6000-9500 SnGb, SnBr, SnJt consimilis open alkaline valleys; Sagebrush Scrub - skirting the deserts from San Diego County north hololeucus sandy neutral soil of high deserts; Alkali Sink, Sagebrush Scrub, Pinyon-Juniper Woodland 3000-9000 west ends of Colorado and Mojave deserts mohavensis well drained, scarcely alkaline soils; Joshua Tree Woodland, Creosote Bush Scrub 2500-6000 w. Mojave desert
Munz gives no elevation range for ssp. consimilis; the Jepson Manual (henceforth, JM) gives 1000 - 2900 m (3300 - 9500 feet) as its elevation range in all of California. A fifth subspecies, leiospermus, is found in the mountains of the east Mojave Desert, and so is not found in the San Gabriel Mountains or nearby. (The word ssp. will be dropped in the rest of the paper for brevity.)
We thought we found three subspecies in this area in two surveys in September and October 2001. We found and correctly recognized mohavensis and hololeucus. We ascribed all the remaining plants to "bernardinus". We did this for a number of reasons:
- First, the plants all looked pretty similar, with the primary difference being that some had whitish stems and some had greenish stems.
However, we were puzzled by the green-stem form of "bernardinus" that we found. The floras state that bernardinus ranges from a white-tomentose form to a green form, so the fact that we found plants at both extremes did not bother us. The two forms were found intermixed at all locations in which either one occurred, which was geographically distinct from the hololeucus and mohavensis populations. But the stem of the green form appeared to be glabrous, not tomentose or felted, and hence couldn't be keyed even to Chrysothamnus nauseosus. However, there was no other possible species identification (henceforth id) for these plants, so we knew they had to be C. nauseosus. And because they clearly looked very similar to the white-tomentose form, and the floras stated that bernardinus had a green form, that id seemed fairly clear.
- Second, they fit the overall characteristics given in the floras for bernardinus reasonably well, at least in our interpretation of the floras at the time.
- Third, all these plants were at elevations of ~6000 feet and above, in Montane Coniferous Forest, which fit the habitat and elevation range of bernardinus bang-on. We don't think we even entertained the possibility that some of these plants might be consimilis, since we don't think of the Montane Coniferous Forest as containing open alkaline valleys; Sagebrush Scrub.
Three years later, we finally learned that the white-tomentose form was indeed bernardinus, but the green form was actually consimilis. Looking back, we realize now the fatal effect of our assumption that we did not expect to find consimilis there. At the time, we had not yet learned to key out plants, and like many amateurs, relied on other characteristics in the descriptions to help us identify plants. In this case, as mentioned above, we used the habitat given in the floras to rule out consimilis, since none of these plants were in open alkaline valleys or Sagebrush Scrub. Nor were these plants occasional, as given in Munz for consimilis; there were tons of them. In any case, as detailed below, the key is ambiguous enough so that we may still have been misled on the id if we had tried to key them out.
As a result, for three years we were somewhat confused by bernardinus, and yearned to see consimilis for comparison. Several times, as we were doing plant lists, and we learned how to key out plants, we found plants that looked like "bernardinus", as we thought we knew it, yet they would key to consimilis. But none of these plants were ever in open alkaline valleys or Sagebrush Scrub, some were definitely not skirting the deserts, and they looked (to us) just like plants that keyed to bernardinus perfectly fine.
In 2003, Tom made two special trips solely to find some certified consimilis. Beauchamp, in his Flora of San Diego County, listed two locations for it: the Narrows, in Anza-Borrego Desert State Park, and Descanso Junction, at the south end of the Cuyamaca / Laguna Mountains.
In the heat of late summer / early Fall, on September 23, Tom surveyed along SR76 to S22, Borrego Springs Road, and SR78 in Anza-Borrego Desert State Park, and back on SR78 to Julian, including a hike in temperatures of 100 degrees along The Narrows Earth Trail and beyond. Tom found plenty of yellow-flowering bushes on this expedition, but none of them were rabbitbrush.
Six days later, Tom visited the second location, with a friend who knew exactly where the plants used to be. We both surveyed the area closely, but found no plants in bloom. We did find one possible specimen, but it was largely inaccessible, and was not of interest yet anyway since it wasn't in bloom.
Finally, on October 21, doing a survey along SR79, SR371 and SR74 en route from Temecula to a hike above Idyllwild, jackpot! Tom found plants that had to be consimilis where SR371 crossed the Thomas Mountains.
This location fit the habitat given in the floras, a dry location next to an alkaline valley, east of the community of Anza. The plants looked bang-on for consimilis, with a distinctly paniculate inflorescence. It fit the properties of consimilis given in Munz well, and because of the location, all other possible ids were ruled out.
Furthermore, there was a distinct twist to the phyllaries, which we had never seen before, nor was it mentioned in any flora. That is, instead of the vertical rows of phyllaries seen in the other subspecies, the rows of phyllaries were at an angle of perhaps 30-40° from vertical.
The stems still posed a problem, since they appeared to be glabrous, just like the green-stemmed version of "bernardinus" we had seen in the San Gabriel Mountains. However, playing with the stem revealed that the outer layer of the stem could be punctured and then peeled back. The outer layer appeared to consist of tomentose hairs that were embedded in some sort of matrix, which effectively made the stem appear glabrous, with no hairs visible as such on the stem.
We now thought we had all these subspecies understood, both for their botanic characteristics and their geographic distribution in Southern California.
We were secure in that knowledge until Tom hiked the Devil's Backbone Trail to Mt. Baldy on August 14, 2004, and made a plant list from memory, obtaining the specific ids by using a list of vouchers kindly provided by the Rancho Santa Ana Botanic Garden Herbarium. When Tom saw the voucher record for consimilis for the area, he could hardly believe his eyes.
We then immediately returned to the Mt. Baldy road area to study the rabbitbrush there. To our surprise, it was clear consimilis.
That meant all of our "bernardinus" ids along the Angeles Crest Highway were in doubt, so we resurveyed that area on August 27, 2004. We began our survey on the coastal side, and were surprised that initially all we saw was consimilis. When we came across our first specimen of true bernardinus, we recognized it immediately as being different, and finally everything fell into place.
In fact, it is quite easy to distinguish consimilis from bernardinus by the appearance, especially using the width of the leaves, the overall leafiness of the stem, and the stem color. However, as is often the case, you have to know what to look for, and you have to have seen both taxa. These characteristics can be seen in pictures of both species, and are discussed below.
In this case, our primary difficulty that we had in distinguishing these taxa is a perfect example of how hard it is to use keys if you haven't seen both taxa. The JM key to distinguish bernardinus from consimilis is:23. Lvs 1-3 mm wide, 1-5-veined; involucre > 10 mm ....ssp. bernardinus
23'. Lvs 1 mm wide or less, 1-nerved; involucre < 10 mm .... ssp. consimilis
- It turns out that many plants of both species have an involucre that is ~10 mm long, which makes that part of the key useless for those plants.
- Every leaf from the many specimens of both taxa that we have seen in Southern California has always been 1-veined; we have never seen a leaf from any true bernardinus with more than 1-vein. So that characteristic is useless.
- Leaves from consimilis are typically ~1 mm wide in the upper stem, but we have seen many plants whose largest leaf anywhere on the plant is ~2 mm wide. So until we had seen and recognized a true bernardinus, we couldn't use this key to produce a consimilis id, since we had some leaves wider than 1 mm.
However, once we saw that the typical leaf on bernardinus plants has a width of ~2 mm, and that the typical leaf on consimilis has a width of ~1 mm, suddenly the leaf portion of the key is obvious, and very reliable.
This once again proves the maxim that in order to understand a key, one often has to see both species.
Date Road Portion Surveyed 21 September 2001 Lone Pine Canyon Road entire length Angeles Crest Highway (SR2) miles 59 to 50 (Lone Pine Canyon Road to Vincent Gap) Big Rock Creek Road (4N11) Vincent Gap to Valyermo Road Big Pines Highway Big Rock Creek Road to Largo Vista Largo Vista Road first mile from Big Pines Highway 11 October 2001 Lone Pine Canyon Road entire length Angeles Crest Highway (SR2) miles 59 to 55 (Wrightwood to Big Pines Highway) Big Pines Highway SR2 to Largo Vista Road 19 August 2004 Mt. Baldy Road entire length 27 August 2004 Angeles Crest Highway (SR2) miles 0-57 (La Cañada Flintridge to Wrightwood)
The sampling interval was ~1 mile for all dates except for 27 August 2004, for which it was ~4 miles. Mileages are referenced to locations in our Angeles Crest Highway (SR2) Road Guide.
All physical characteristics of each specimen were measured in the field in 2001. In 2004, the leaf, involucre and corolla properties were measured using a dissecting microscope; all other characteristics were measured in the field.
Vouchers were obtained on the 27 August 2004 fieldtrip which will be deposited at the Rancho Santa Ana Botanic Garden Herbarium.
III. Botanical Properties
This section contains a discussion of how to separate the species, and gives a key to do so. Many of the properties summarized here are explored in detail in subsequent sections.
As mentioned above, we found it easy to immediately identify hololeucus and mohavensis, but did not distinguish bernardinus from consimilis in our 2001 work. Thus in some of the plots below, the latter two species will be combined.
The white form of hololeucus is by far the most easily recognizable, even from a significant distance: it looks like a white ghost and stands out dramatically from its surroundings, due to its white stems and white leaves. However, hololeucus comes in a greenish form as well, which is also visible in the linked picture, which must be examined up close to identify it.
All hololeucus are easily identified by their erect tiny corolla lobes; all other subspecies have spreading corolla lobes. That is, you simply look at any of the five small flowers within each flower head. The small flowers of all subspecies except hololeucus look like a typical flower, with five lobes angled at 90° from the corolla tube, allowing you to see the full surface area of each lobe when you are looking into the flower. In contrast, the lobes for hololeucus look like a continuation of the corolla tube with slits on its side. As a result, the flowers of hololeucus are significantly less showy than those of the other species, since you see only the edges of the corolla lobes when you are looking into the flower. (pictures to be supplied)
Our hardest challenge was separating consimilis from mohavensis. We could find no reliable way to distinguish these two species in the San Gabriel Mountains, using any of the properties that are said to distinguish them in the floras.
It finally became apparent to us that these two taxa form an intergrading population in the San Gabriel Mountains, as shown clearly by the following plots.
The floras separate these two taxa by the length of the involucre, the length of the leaves, and by whether the stem is leafless or leafy at flowering time. (add table with values from floras). We find no correlation at all of any of these parameters with each other. We will analyze the correlation of these characteristics in the new set of plots.
We measured the maximum leaf lengths for each plant, and plotted them versus the median involucre length:
(add boxes from floras). As can be seen, there is no correlation at all. This frustrated us immensely, since the floras indicate that consimilis has leaves 10-50 mm long, and mohavensis has leaves 10-30 mm long.
The JM key also distinguishes the two species by whether the stem is leafless or not at full bloom. We found no correlation at all with involucre length:
The histogram of all involucre lengths shows no apparent break point around the 9 mm separation point in Munz:
The peaks at integer lengths are due to some measurements that were made only to the nearest mm. If one smooths the curves by distributing some of the integer measurements into neighboring bins, this histogram is not very different from a smooth Gaussian-like distribution. That is, there is no evidence for the existence of two separate taxa in this population.
We initially tried to classify all plants as either mohavensis or consimilis by imposing a break point near 8.25 mm in the median involucre length for each plant or set of plants at each location. However, there were a significant number of plants that were at the break point, implying this division was artificial, and we could not find any other characteristic that supported this division. A break point at any other value worked no better.
Nonetheless, we continued to cling to the belief in two separate taxa until we made the following plots.
A plot of the median involucre lengths vs. latitude and longitude shows a smooth variation from east to west and north to south:
Even with this breakdown with geography, there is no clear break point in the involucre lengths near 9 mm. Seeing this trend with geography immediately dispelled our belief in two separate taxa, since the smooth variation with geography clearly implies the two taxa intergrade in this area. The JM mentions that these two species sometimes integrade; there is no better evidence for an intergrade than the above plots.
A geographic breakdown by involucre length clearly shows the trend:
The other two subspecies are also easily separated by their leaves. The leaves of consimilis are mostly threadlike (with their width ~equal to their thickness), with typical widths of ~1 mm, and are not an obvious feature of the plant. In contrast, bernardinus appears leafy, with leaves distinctly wider than thick, and typical widths of ~2 mm. See pictures of the plants and a flower stem of consimilis next to one of bernardinus.
In addition, plants of consimilis are significantly taller than plants of bernardinus. These plants are often found side by side, and this property makes it easy to distinguish them from a distance. However, because the plants vary in size with age, it is not possible to place a numeric value on this, except for the oldest, tallest plants.
Here is a key to distinguish the subspecies:1. Corolla lobes erect; plant often appearing as a "white ghost", with white leaves and stems.... hololeucus
1' Corolla lobes spreading; plant never appearing as a "white ghost", with greenish leaves, with or without a white stem2. Stems leafless, or with mostly dried leaves, at time of flowering; involucre sharply angled....mohavensis
2'. Stems gen with most leaves still living at time of flowering; involucre angled, but not sharply so3. Leaves thread-like, typical height and width ~ 1 mm; stems white-green or yellow-green, without visible hairs on the stem surface; plants tall ... consimilis
3'. Leaves wider than thick, typical width ~2 mm; stems gen white, usually with visible white hairs on the surface; plants short ... bernardinus
As discussed in the introduction, the stems of consimilis do not appear to have any hairs on them at all to the eye. But there is actually an outer layer on the stem that contains whitish hairs in some sort of matrix. If you scrape off the outer layer under a microscope, you can see the ends of the hairs protruding from the cut sections.
The color of the stems depends on the eyes of the person viewing the stem, and probably the lighting conditions. The color of the stems of consimilis is said to be yellowish-green in the floras. Jane at times can detect a yellowish cast to the stems. Tom always sees the stems as white-green, as if a green stem had been white-washed with a thin coating of milk. Each of us sees the stems of bernardinus as white. Note that there is supposed to be a green-stemmed version of bernardinus, but we have not seen any so far.
The subspecies can usually also be easily distinguished in the field by their bloom stages (see next section).
Bloom Times And Dependence On Elevation
Each of the subspecies have different elevation-dependent bloom periods. The following three plots show the bloom stage of each species on the three times we surveyed large numbers of plants of each species.
Consider first the 2001 data:
21 September 2001:
11 October 2001:
These data are all from the desert side of the mountains. There is a clear trend for each species to begin blooming at the highest elevations, and for the bloom to be progressively delayed with decreasing elevation.
Hence on 21 September 2001, essentially only bernardinus and consimilis were in bloom, and they were in full bloom or past full bloom. Only a single plant of hololeucus was in full bloom; all other plants of both hololeucus and mohavensis were only in bud or had just a few flowers.
By 11 October 2001, every plant of consimilis and bernardinus was past full bloom, with half of the plants nearly completely finished blooming. In contrast, all plants of mohavensis and hololeucus were now in full bloom at most elevations, with only the highest elevation plants past full bloom.
27 August 2004:
IV. Geographic Distribution
Add table summarizing properties of each species; i.e., check a column for each species if it has a given trait (st color, invol length, erect corolla lobes, etc.)
Also mention lepidospartum spp and senecio douglasii.
Copyright © 2004 by Tom Chester and Jane Strong
Permission is freely granted to reproduce any or all of this page as long as credit is given to us at this source:
Comments and feedback: Tom Chester | Jane Strong | Michael Charters
Last update: 13 September 2004