Plants of Southern California: Gutierrezia californica and G. sarothrae: Initial Investigation

This page reports on my initial investigation into these two species. See the latest summary page for later work on these two species, including how to separate them, how well they separate in southern California, and geographic maps of their distribution in southern California.

Table of Contents

     Current understanding of these two species
     Motivation for this study
     Results for First Two Populations
     Results for First Six Populations
         Involucre Height and Width
         Number of flowers
         Clustering of heads


Current understanding of these two species

Gutierrezia californica and G. sarothrae are two species in what Lane 1985 says has been called the "G. sarothrae complex", which contains five species that differ very little in qualitative characters; most interspecific differences are quantitative (Lane 1985, Taxonomy of Gutierrezia in North America, Systemic Botany, 10(1), p. 8).

Worse, Lane wrote that these two species intergrade in southern California, which if true, would make it impossible to determine some specimens as one species or the other. However, Solbrig 1965, The California Species of Gutierrezia, stated that:

Viewed from the vantage point of good knowledge of the species in the field, and after carefully analyzing a number of breeding populations, the species appear fairly distinct. The same cannot be said, unfortunately, when the problem at hand is the identification of herbarium species [since] large specimens of G. sarothrae and small ones of G. bracteata [=G. californica in Lane 1985] are likely to be confused.

In this analysis, I include G. bracteata within G. californica, following Lane.

The difficulty of the separation of individual specimens of these two species in southern California is such that there are 17 duplicate vouchers that have one voucher called G. californica and the other voucher called G. sarothrae. There is even one set of duplicate vouchers, Solbrig 2755, with one voucher (UC1214009) determined as G. californica by Solbrig, with the other voucher (RSA152485) redetermined as G. sarothrae by Lane! This particular voucher is of a population that has a chromosome number and involucre size consistent with G. californica, but the mean number of flowers is very low, more typical of G. sarothrae.

Both Lane and Solbrig were expert in the Gutierrezia genus, and their differing conclusions may have resulted from their different approaches to studying this genus, as well as perhaps from each of them emphasizing different characteristics as being more important to separate the species. Solbrig studied twenty populations of Gutierrezia in California, measured up to nine characters in 50 plants in each population, and analyzed the difference in the means of each of those characters. He also concentrated on chromosome numbers, finding that all plants of G. sarothrae in California had n=4, with plants of G. californica having n=8 and n=12.

Lane studied the plants over a much larger geographic range, 14,869 herbarium specimens and ~1500 specimens collected during several field trips through most of the ranges of all taxa. One of her major conclusions was that G. sarothrae was an exceedingly variable species:

There is a high degree of environmentally induced phenotypic plasticity within populations and even within individual plants on a season-to-season basis. Peduncle length is also affected by moisture levels, as are numbers of heads and florets per head. Edaphic factors also readily affect phenotype in this species; common effects include an increase in number of florets and thus in diameter of the capitulum and, often, a concomitant reduction in the number of heads. Such changes are also often noted with an increase in polidy level in this species, but neither firm correlation between edaphic parameters and polyploidy nor reproductive isolation as a result of differences in ploidy level have been demonstrated.

She also found that G. sarothrae has individuals in its full range with n=4, n=8 and n=16, although it may still be true that its California populations all have n=4.

Despite this variation within G. sarothrae that needed more study before she would recognize any varieties within it, she recognized G. californica as a distinct species

because the morphological differences between it and G. sarothrae are correlated with ploidy level and geographic distribution; further, the combination is well-established in the literature.

However, she specifically stated that further work needs to be done to understand how best to treat G. californica, as well as to determine whether G. sarothrae could reliably be divided into varieties.

Lane's maps of the geographic distributions of the species given in Fig. 1 give some insight into why these species might be difficult to separate, especially in southern California. G. sarothrae is a very widespread species throughout the West, whereas the distribution of G. californica looks like its much-smaller range is just a small bite taken out of the G. sarothrae distribution. The two species have an area of overlap in southern California and northern Baja Californica.

Fig. 1. Geographic distribution maps from Lane 1985. Left: Gutierrezia californica. Right: G. sarothrae.

Motivation for this study

As a beginning botanist, not really understanding the situation in southern California for these two species, I took the key in the Jepson Manual as gospel, and ended up naively determining the plants from two neighboring areas as the two different species, even though I now believe both sets of plants are the same species, having only minor differences. It was unfortunate that their minor differences in the clustering of heads tipped one population just over the line to separate them, and the other population just under the line!

A few years later, when I saw a specimen of G. sarothrae in the Mojave Desert, it struck me as being very different from our plants in southern California. I then looked at online pictures of G. californica from the central coast, and they, too, looked very different from our plants. After that, I thought that our plants in southern California were not readily assignable to one species or another, which makes it difficult to give them a name. So for each area for which I did a flora for in southern California, I decided to go with the flow and just give them the name that was most common in vouchers.

Finally, in 2015, I decided I ought to analyze some specimens to see quantitatively what the situation actually was in southern California, whether most plants are intermediates, and/or whether there are clearly defined examples of one or both species here as well. Recall that Solbrig maintained that every population could be uniquely assigned to one of the species, whereas Lane stated that where the ranges overlap, there is significant morphological intergradation.

It will be a while before I have enough data to do a Principal Components Analysis to get the answer to my question. Until then, this page just presents some of the data as I accumulate it.

Measurements So Far

Results for First Two Populations

On 18 September 2015, I sampled plants from two different somewhat nearby populations in Riverside County, from the Cactus Spring / Sawmill Trail parking area in Pinyon Flats, and from near the junction of Morris Ranch Road and SR74 in Garner Valley. Morris Ranch Road is 5.1 road miles west of Santa Rosa Summit along SR74, which separates the desert drainage from the coastal drainage, and the Cactus Spring turnoff on SR74 is 7.9 road miles east of Santa Rosa Summit.

In each population, a single flowering stem was measured on five different plants, since Solbrig (1965) emphasized that to reliably distinguish the species one needed to compute the mean values for a population due to the large variability in the characteristics of individual heads.

Much to my surprise, both sets of plants turn out to be perfect G. sarothrae. Table 1 gives the measured values compared to the means and range for each species given in Solbrig and/or in Lane. Note that Solbrig only had three measured populations of G. sarothrae and nine measured populations of G. californica, and hence the range of his means is almost surely smaller than the true range of the means when more populations are considered.

Table 1.

invol length (mm)invol width (mm)# ray flowers# disk flowerstotal flowers
Cactus Spring
Garner Valley
G. californica from Solbrig and Lane
rangegen <6.5gen <3.54-134-138-20
G. sarothrae from Solbrig and Lane
rangegen <4.5gen <2.52-82-96-14


Taken together, the evidence is very strong that both the Garner Valley and Cactus Spring plants fit the floral keys and descriptions for G. sarothrae, and not G. californica, without any evidence for intermediates.

Fig. 2 gives separate histograms of the number of both types of flowers for each area.

Fig. 2. Histograms of the number of ray flowers, and the number of disk flowers, for all measured heads from Garner Valley (top) and Cactus Spring (bottom).

I also analyzed the clustering of the flowers for these samples, but I defer the discussion of this to later in this analysis.

Results for First Six Populations

Involucre Height and Width

On 23 September 2015, I sampled plants from four populations in San Diego County: Ranchita, Scissors Crossing (where the plants had been vouchered as G. californica); Inspiration Point just south of Julian; and at the Lake of the Woods on the Sunset Trail in the Laguna Mountains. Surprisingly, all populations were clearly G. sarothrae.

Fig. 3 presents the mean involucre height and width for these four San Diego County populations and the two Riverside County populations, along with the data given in Solbrig 1965. The six populations clearly fall within the range for the three populations of G. sarothrae measured by Solbrig, far from the range of his measurements for G. californica.

Fig. 3. Mean involucre height vs. mean involucre width

Number of flowers

Fig. 4 presents the mean number of ray and disk flowers, again with Solbrig's data. Once again, the six populations fall within the range for G. sarothrae, although Solbrig has one sample of G. californica that has even fewer ray and disk flowers. Interesting, this sample is the population discussed in the introduction where Lane determined one of Solbrig's dupe voucher as G. sarothrae!

Fig. 4. Mean number of ray flowers vs. mean number of disk flowers

Fig. 5 gives a histogram of the number of ray flowers per head, and of the number of disk flowers per head, for 113 heads from the 30 samples (five stems from each of six populations). Fig. 6 shows the histogram of the total number of flowers per head. All ranges are dead consistent with the ranges given by Lane for G. sarothrae, and inconsistent with the ranges for G. californica. The apparent dip in the histogram at 12 total number of flowers is not statistically significant, being a fluctuation of just 1.7 standard deviations. Such a fluctuation due to chance would be expected for at least one point in such a histogram.

Fig. 5. Histogram of the number of ray flowers per head, and of the number of disk flowers per head

Fig. 6. Histogram of the total number of flowers per head. The dip at total number of flowers = 12 is not statistically significant.

For completeness, Fig. 7 gives a histogram of the difference in the # ray and disk flowers in each head. The majority of the heads have the number of ray and disk flowers the same, or differing by just one flower.

Fig. 7. Histogram of the number of ray flowers minus the number of disk flowers in each head.

Clustering of heads

This characteristic is poorly defined, and extremely hard to interpret in specimens. Yet it is probably the single characteristic used for specimen determination, since the range in the number of flowers per head is difficult and tedious to measure in vouchers (the Jepson Manual key uses only the number of flowers and clustering of heads to separate these two species). Hence I suspect that how one interprets the clustering of heads is a prime source of the differences in determinations for duplicate vouchers.

Fig. 8 shows one of the major difficulties in using the clustering of heads to determine specimens: the clustering of heads varies from the lower to upper inflorescence branches, and from plant to plant. Taken alone as a voucher specimen, I suspect the rightmost sample in Fig. 8 would be determined as "G. californica"; the leftmost two samples would be determined as G. sarothrae; and the other sample might be called an "intermediate". Yet the five specimens from Garner Valley (one is not shown in Fig. 8) are perfect G. sarothrae for the involucre heights and widths, and numbers of ray, disk and total flowers!

Fig. 8. Four samples of G. sarothrae from Garner Valley, in random order of being pulled from my sample bag. The rightmost sample has mostly-single heads on its ultimate inflorescence branches; the leftmost sample has mostly-clustered heads.
Click on the picture for a larger version.

Another example is from the plants at Scissors Crossing, which have been vouchered only as "G. californica" even though they are perfect G. sarothrae in their involucre and flower characteristics. It was immediately obvious in looking at the samples that the heads were mostly single, on long peduncles, which would key them to G. californica. Yet one of these samples also has clusters of 4 heads.

Fig. 9 shows another difficulty in interpreting whether the heads are clustered or not: how close do the heads have to be to qualify as being "clustered"? There is no easy answer to this, since there is a continuous variation in how close together heads are at the tip of the ultimate inflorescence branches. (See also photograph of all five samples from Inspiration Point.)

Fig. 9. Tips of one inflorescence from a single specimen of G. sarothrae from Inspiration Point just south of Julian. Everyone would probably agree that group #1 has four heads (one head is nearly hidden behind another), and nearly everyone might agree that group #3 has either a cluster of three heads, or a cluster of two heads with a nearby single head. But group #2 is far more open; some people might consider this a cluster of three heads, while others might consider it to be three non-clustered heads. Group #4 is more open yet; most people might say these are not a cluster at all, and instead are made of three different clusters of heads, with one head at left; two heads in the middle; and a cluster of three head on the right.
Click on the picture for a larger version without the labels.

Yet another difficulty with this characteristic is exactly what does it mean to say, as in the Jepson Manual key: "heads generally 1" or "heads generally in clusters of 2-5"? As I quickly discovered as a beginning botanist , it makes a big difference in this couplet in how you interpret the key. To make this clear, considered the following example which roughly corresponds to our actual situation in southern California. Suppose a plant has 60% of its branchlets with a single head and 40% with two heads. Then 60% of the branchlets, the majority, have a single head. But 57% of all the heads are found in clusters of two. So does the key element refer to the number of branchlets, which would then determine this plant as G. californica, or the number of heads, which would then determine this plant as G. sarothrae?

A fourth difficulty is whether immature heads should be counted in a grouping. The specimens I have looked as so far often have immature heads in the inflorescence, which may be heads that will never mature, or may be heads just waiting their turn to mature.

By now it should be clear that at least for our southern California plants, the "clustering of heads" is a terrible characteristic to use to determine the species.

Nonetheless, to see how the numbers played out for these samples, I attempted to characterize the grouping of the heads at the tips of every main inflorescence branch in every sample. I'm sure my subjective assessment changed with time due to these four uncertainties in how to count the clusters. Sometimes I would take the dominant grouping and report that as the number of heads per cluster at each inflorescence branch tip. Sometimes I would take the maximum number of heads in a group if that was clear, since only G. sarothrae was supposed to have clusters of 4 or more heads. I would usually count immature heads if the inflorescence was young, but sometimes I ignored them if the inflorescence was on its last flowers. I usually considered a cluster to be when there was overlap in the involucres and the involucres were no farther apart than perhaps twice the involucre width. Specifically, in Fig. 9, I probably usually took Group 1 as a cluster of 3 heads; Group 3 as a cluster of two heads, and Group 2 as non-clustered heads.

Fig. 10 shows the results of that assessment; the percent of inflorescence branch tips with a single head is all over the place, from 10 to 70%, despite these plants all being clear G. sarothrae in the characteristics of their involucre and number of flowers!

It is interesting that the two samples with over 50% of their inflorescence branches having single heads at their tip are determined differently in vouchers. The Scissors Crossing plants are determined as G. californica, whereas the essentially-identical plants from Garner Valley are determined as G. sarothrae.

Fig. 10. The percent of inflorescence branch tips judged to have heads single, or heads in clusters of 2 to 5, separated by area. The number of inflorescence branches ranged from 31 for Garner Valley to 44 for Lagunas Sunset Trail.

This is the end of my current analysis. My goal now is to try to find some real G. californica to get a proper comparison with these G. sarothrae plants.

Go to Native and Introduced Plants of Southern California

Copyright © 2015 by Tom Chester
Permission is freely granted to reproduce any or all of this page as long as credit is given to me at this source:
Comments and feedback: Tom Chester
Last update: 4 October 2015