Plants of Southern California: Gutierrezia californica and G. sarothrae
Table of Contents
Summary
Introduction
The Gutierrezia californica and G. sarothrae species complex
Distinguishing the two species
The mean involucre height and width
The mean number of ray and disk flowers
The "intergradation" of the species in southern California
Geographic Distribution
Analysis of vouchers for San Diego County
| 
| |
| Fig. 1. Left photo: The two heads on the left are G. californica; the two heads on the right are G. sarothrae heads randomly selected from the populations with the shortest involucres.
Right photo: The heads on the top are G. sarothrae, arranged in increasing order of mean involucre size for the population, but with a random selection from each population. Note that even though the heads on the right come from the population of G. sarothrae with the largest mean involucre length, the involucres on these randomly-selected heads are shorter than the head to the left of it. The heads on the bottom are G. californica. Note that the middle G. sarothrae head is essentially identical in shape and overall appearance to the larger G. californica head below it. Each photo has mm tickmarks shown on the ruler at top. Click on the pictures for larger versions that have each flower head labeled. | ||
Summary
- Gutierrezia californica and G. sarothrae are two different species, with different chromosome numbers, that cannot interbreed in normal breeding, and therefore cannot form hybrids.
- These two species are very similar since G. californica was derived from G. sarothrae by chromosome doubling and tripling, in a rare event sometime in the past.
- The range of variation in these two species is such that they can be easily separated at their morphological extremes, but a significant number of specimens of each species are difficult to distinguish morphologically from the other species.
- Populations of each species can be more easily distinguished morphologically than can individual specimens. The mean involucre height and width, measured from at least 15 heads from five different individuals in a population, appears to discriminate the species with a reliability better than 90%.
- The population mean number of ray and disk flowers per head only generally separates the species in southern California, with a reliability of significantly less than 90%, and is completely misleading in some populations.
- The difference between the two species in the number of ray and disk flowers appears to be significantly less in southern California than in areas outside southern California.
- The "clustering of heads" is a poorly-defined characteristic, extremely hard to interpret in specimens, and is quite variable even within a single inflorescence from a single plant. It is an unreliable separator of the species in southern California.
- The previous two characteristics probably are the basis behind the statement that the two species "intergrade" in southern California.
- Previous voucher determinations of plants in southern California are extremely unreliable, producing incorrect maps of the geographic distribution of each species there. This determination problem probably results from the difficulty of determining an individual plant; the more subtle differences between the two species in southern California; and the poor keys in the floras to try to separate these species here.
- Using population measurements, geographic maps are derived of the distribution of each species in southern California.
Introduction
For introductory material to this investigation, see Gutierrezia californica and G. sarothrae: Initial Investigation.
The Gutierrezia californica and G. sarothrae species complex
G. californica was derived from G. sarothrae (n=4) by chromosome doubling (n=8), in a rare event sometime in the past. The G. californica plants with chromosome tripling (n=12) either came independently from G. sarothrae alone, or via a crossing of G. californica and G. sarothrae, again in a rare event in the past.
Thus the n=8 plants of G. californica started out with two copies of the G. sarothrae genome. Sometimes in such cases when the number of chromosomes doubles, significant alteration in some of the genes occurs in the first several generations until the new genome develops into one that is stable in reproduction. This may or may not lead to immediate significant differences between the tetraploid and the diploid. If not, the doubled genes may be preserved mostly intact, and then only slowly change with time as the species go their separate ways.
Even though this process of gene doubling in progeny is rare in any given reproduction event, it occurs quite often in plant species over time, and is called polyploidy. Between 30% and 70% of angiosperm species are believed to be of polyploid ancestry, including polyploids derived from gene duplication in a single species and ones derived from crosses of two different species. Hence the origin of G. californica is a quite common way in which species originate.
Each type of polyploid has a different "-ploid" label. The species with the original number of chromosomes, G. sarothrae in this case, is called a diploid species, since in most of its cells it contains two copies of each chromosome, one from each parent. If the number of chromosomes is doubled, that species is called a tetraploid, having four copies of each chromosome. If the number of chromosomes is tripled from the diploid species, that species is called a hexaploid, with six copies of each chromosome.
The extra copies of genes often produce more proteins, etc. to help them grow better and cope with environmental challenges. When that happens, tetraploid species often are simply somewhat-more robust versions of the diploid species, somewhat larger in plant heights, flowers, etc. As they evolve in time, tetraploid species have an advantage in adapting to different environmental conditions, since they can use one copy of the genome to "experiment" with different random genetic changes, and still keep one copy of the genome to keep things going well for the functions those chromosomes performed.
In most cases, species with an increased number of chromosomes are difficult to distinguish from the original diploid species. In those cases, botanists lump the polyploids together under one species name. One example is Larrea tridentata, creosote bush, which has diploid populations in the Chihuahuan Desert; tetraploid populations in the eastern and southern Sonoran Desert; and hexaploid populations in the northern and western Sonoran Desert, and Mojave Desert (see the geographic distribution map, Fig. 2 of Phylogeny and Cytogeography of the North American Creosote Bush (Larrea tridentata, Zygophyllaceae), by Laport, Minckley and Ramsey 2012).
In some cases, as for Gutierrezia californica and G. sarothrae , the tetraploids (and hexaploids) are at least somewhat morphologically distinguishable from the diploids, and are given two separate species names.
If the situation is intermediate between these two cases, with the species being distinguishable morphologically at the extremes but indistinguishable in other cases, the group is often called a species complex, leading to differences of opinion as to whether the tetraploids and diploids should be lumped into one species or split into two species. Lane 1985 considers these two species to be part of a species complex, along with two other similar Gutierrezia species that are easier to separate from these two species, but still ended up treating these two species as separate species since their morphological differences correlated with ploidy level and geographic distribution.
Gutierrezia californica and G. sarothrae are an example of both lumping and splitting. The tetraploids and hexaploids (n=8 and n=12) have been lumped under the name G. californica, but separated from the diploid G. sarothrae.
Distinguishing the two species
The only 100% reliable way to separate these two species in southern California is to count their chromosomes! Morphological measurements can separate populations with fairly high reliability, perhaps something like 90% or higher. However, such measurements are less reliable in determining individual specimens.
The major difference between the two species is of size. Since the size of an individual plant also depends on moisture conditions, it is difficult to morphologically differentiate the two species at times. Gutierrezia californica growing under poor conditions is essentially indistinguishable from G. sarothrae growing under good conditions.
Worse, the concept of the two species (what characteristics delineate them), is different in each of the major floras for California. The keys in the major floras do not reliably separate the species in southern California, since they use characters that do not reliably separate the species here.
It is therefore not surprising the vouchers are hard to determine properly, with people disagreeing over how to determine a given voucher. 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.
Due to the variance in species concepts for these two species, and with doubts expressed by botanists about whether these two species were validly different, Solbrig undertook a study in 1965 (Madrono 18, p. 75) to study these two species by intensive field studies and population analysis.
Solbrig's insight was that if one measured a large number of plants (50!) in a number of different areas (20), one could get a better measure of the intrinsic properties of each species, averaging out variation at a site due to moisture conditions, as well as specific genetic variants within a population.
His conclusion was that the species appear fairly distinct in population studies, but not so much in individual herbarium specimens. In particular, large specimens of G. sarothrae and small ones of [G. californica] are likely to be confused.
I have undertaken field studies in 2015 to see how well the two species separate when one measures a single inflorescence from five separate plants at a given location, and averages the results. My results are in good accordance with Solbrig's, and it appears to be possible to make species determinations with better than 90% reliability using such measurements.
However, averaging five plants at a given location in a given year will not average out year to year variation. I plan to study herbarium specimens to see if averaging specimens geographically will produce the same reliability in distinguishing the species. Results so far appear to indicate that the two species are mostly separate geographically, but that needs to be tested with measurements from different years.
The mean involucre height and width
Fig. 1, at the top of the page, shows the involucres from six different populations, all in the same photograph to show the relative sizes, along with a ruler to give the scale. It is readily apparent that the involucres for the two heads shown for G. californica are significantly taller than the involucres for the other heads shown for G. sarothrae.
However, it is nearly impossible to distinguish the two species in photographs without a scale. Fig. 2 shows the involucres from every measured population, but now without any scale in each pix. I have tried to scale each involucre to the maximum size in the photos in Fig. 2, but some of the pictures did not have enough resolution to make the displayed involucres the same size. The pictures are thus not to the same scale.
Two of the photographs in Fig. 2 show G. californica; the rest show G. sarothrae. Can you pick out the two photographs of G. californica? Keep your mouse away from the photographs if you don't want to see the determination given at the bottom of your browser window.
I cannot pick out which ones are G. californica in this gallery, even though these are my photographs. If you can, I would love to know what you saw in those two photographs to distinguish them from the others. You can discover which two photographs are of G. californica by looking at the URL of the linked photograph, or by looking at the bottom of this page.
| 
| 
| 
|
| 
| 
| 
|
| 
| 
| 
|
| Fig. 2. Photographs of the two species that are not to the same scale. I have taken my best resolution photographs of the flower heads from each location, and attempted to scale each so that the head fills the vertical part of the thumbnail. Some of the locations did not have high enough resolution photographs to fill the vertical length of the thumbnails.
Two of these photographs show the longer involucres of G. californica. Can you pick out which ones they are? See the text for the answer. Click on the pictures for larger versions. | |||
Fig. 3 shows the mean involucre height and width for my population measurements and those of Solbrig.
Fig. 3. The mean involucre height for each population vs. the mean involucre width, for data from Solbrig and from Chester. Note that the data sets from each source correspond quite well, and that both have good separation of the two species. The five Solbrig points with involucre widths greater than 3.0 mm are all from northern California.
The mean values in Fig. 3 separate the two species cleanly. Furthermore, the separation in Fig. 3 correlates precisely with geographic distribution, confirming that these are two different species. Individual head measurements in my data also generally separate the species, but with an accuracy well under 100%. Also, the measurements from individual plants tend to be highly correlated, so averaging heads in a single voucher specimen does not result in much improvement in separating the species.
If you would like to check that you are measuring the involucre length correctly, this version of the photograph linked in Fig. 1 on the right shows where I measured the involucre lengths in that photograph, and gives the involucre lengths I deduce from that photograph. To measure the length of the involucre, one measures from the tippy-top of the involucre to the bottom. One sometimes has to look for the farthest extent of the phyllaries at the top, and sometimes there is a bit of judgment involved with where the bottom of the head is, especially if the bottom of the head is hidden by bracts. Of course, measurements in a photograph are not as accurate as measurements under the microscope, since if the scale is not at the same distance from the camera lens as the head, the heads are not exactly to the scale seen in the photograph.
The mean number of ray and disk flowers
The mean number of ray and disk flowers generally separates the species, but with an accuracy well under 100% in the entire data set, and much less than 100% for the southern California populations. Fig. 4 shows the mean number of ray and disk flowers for my population measurements and those of Solbrig. The only populations that separate out well, with a mean number of disk flowers larger than eight, are G. californica populations from northern California.
Fig. 4. The mean number of ray flowers for each population vs. the mean number of disk flowers, for data from Solbrig and from Chester. The mean number gives a general indication of the species, but deductions from it can be very wrong at times. Note in particular that one Solbrig G. californica population matches the lowest value for any G. sarothrae population, even though in general low values almost always correspond to populations of G. sarothrae. The five most extreme G. californica points are again all from northern California.
The "intergradation" of the species in southern California
Lane 1985 wrote that where the ranges [of these two species] overlap, there is significant morphological intergradation. My data show that the species are closer together in the number of disk and ray flowers in southern California than elsewhere, and possibly that the clustering of heads is more similar in southern California than elsewhere. The data on the number of disk and ray flowers is presented below. The clustering of heads has been discussed previously.
Note that just because the species are closer together morphologically in southern California does not mean that they hybridize, or are impossible to separate. As mentioned before, they cannot hybridize under normal conditions due to their difference in number of chromosomes. The species can still be distinguished in southern California; it just is harder to separate them here.
Fig. 4 showed that the populations of G. californica in southern California have fewer ray and disk flowers than populations of G. californica in northern California. Fig. 5 shows that it is also true that populations of G. sarothrae in southern California have more ray and disk flowers than populations of G. sarothrae elsewhere.
Fig. 5. The mean number of ray flowers for each population vs. the mean number of disk flowers, for data from Solbrig and from Chester, as in Fig. 4, now with the gen ranges for the number of flowers for each species taken from Lane 1985 plotted as rectangles. Note that essentially none of the G. sarothrae specimens in southern California fit the gen 3-5 range for the number of disk flowers, and the number of ray flowers, given by Lane for G. sarothrae. Further, essentially none of the G. californica specimens in southern California fit Lane's gen 6-13 range for the number of disk flowers, and the number of ray flowers, for G. californica.
One hypothesis to account for the decreased morphological separation in southern California is that G. californica first appeared here. As the habitats in the coast ranges of California became available in more recent geologic times, the tetraploid species of G. californica was able to occupy that different habitat, whereas the diploid species of G. sarothrae was not able to do so. As G. californica spread north, it may well have undergone evolutionary selection as it traveled. The populations left behind in southern California may not have been under such evolutionary selection, since they remained in at least somewhat-similar habitats as the G. sarothrae here.
One also needs to explain why populations of G. sarothrae in southern California are a bit more like G. californica here than populations of G. sarothrae elsewhere. It could just be that the relative morphological convergence is due to both species growing in somewhat-similar habitats in southern California, where the environment selects for that middle ground of traits.
Geographic Distribution
Using the mean involucre length shown in Fig. 3, I have assigned the populations I've studied to the appropriate species. I have also georeferenced Solbrig's data in his 1965 paper, and Solbrig's vouchers that are available at the Consortium of California Herbaria. Fig. 6 shows the geographic distribution of those populations in southern California.
Fig. 6. The geographic range of the two species in southern California from the sources detailed in the text. The only place where the two species are found in proximity in this data set is for G. californica in the Palm Springs / Pinyon Flat area, with populations of G. sarothrae immediately to the west (see text below for comment about the southernmost G. californica point possibly being mislocated). Data are sparse in many regions, so there may be additional populations of both species outside the range shown by the solid lines.
I do not know how reliable the determinations are of Solbrig's vouchers from locations where he did not report chromosome determinations or population determinations, nor do I know whether my deductions about the species from the involucre sizes are perfectly correct, since I haven't measured the number of chromosomes. Fig. 7 shows where chromosome measurements have been done by Solbrig, the most rock-solid of the species determinations here.
Fig. 7. The geographic range of the two species in southern California from the sources detailed in the text, as in Fig. 6, now with an overlay of locations where chromosome numbers have been determined. The range of each species from all locations is consistent with the range determined from chromosome numbers.
Fig. 8 plots the elevation of each population vs. longitude.
Fig. 8. Elevation vs. longitude for the two species in southern California from the sources detailed in the text.
The distribution of the two species shown in Figs 4, 5 and 6 looks reasonable, with the exception of the single G. californica point at 4000 feet from the Pinyon Flat area above Palm Desert. This voucher has a chromosome determination, so there is no doubt that it is G. californica, even though those elevations have only populations of G. sarothrae otherwise. The voucher is Solbrig 2774. Oddly, in Solbrig 1965 the location is given as Idyllwild, even though the voucher location says on Hwy 74 18.3 mi e Keen Kamp. Although that location seems pretty precise, it is always possible that an error occurred in Solbrig matching the samples to his field notes. His next collection is the sample of G. californica from near entrance to Tahquitz Canyon at a much lower elevation. That point at 4000 feet should therefore be treated with some skepticism.
In contrast to the mostly-separate geographic distributions of the species shown above, the geographic distribution of G. californica from vouchers in southern California looks almost identical to the geographic distribution of G. sarothrae from vouchers! This is almost surely due to the more subtle differences between the two species in southern California, and the poor keys in the floras used to try to separate these species in southern California. It appears that a large number of voucher determinations are erroneous.
Analysis of vouchers for San Diego County
To check on my speculation about voucher misdeterminations, I visited the herbarium at the San Diego Natural History Museum on 5 November 2015. I measured three involucres from each of the specimens in a folder labeled "G. californica" from San Diego County, and from each of the specimens in a folder labeled "G. sarothrae" from San Diego County. The histogram of the measurements for the involucre lengths was identical for each folder to within measurement error, with the data consistent only with a determination of G. sarothrae for the specimens in both folders. Fig. 9 shows the histogram of the measurements from each folder, along with the mean involucre length for each species from the data presented in Fig. 3 above.
Fig. 9. Histogram of the involucre lengths from individual heads from vouchers determined as "G. californica" from San Diego County, and separately from vouchers determined as "G. sarothrae" from San Diego County. The mean involucre length for each species from my field data is shown as well. The observed two histograms are both consistent only with a determination of G. sarothrae, implying that every single voucher in the "G. californica" folder was a misdetermined G. sarothrae.
The range of measured values for the involucre lengths was the identical 3.0 to 5.0 mm for each folder, with both folders having the identical mean involucre length of 4.0 mm, exactly equal to the mean involucre length of 4.0 mm from my measurements of G. sarothrae populations. Not a single head in either folder came close to the mean involucre length of 5.5 mm from my measurements of two G. californica populations, one from Orange County and one from Cajon Pass.
The locations of these vouchers fit well with the geographic distribution map of G. sarothrae shown in Fig. 6; see the updated geographic range for G. sarothrae in the San Diego County area. I have yet to find a specimen of an actual G. californica from any place in San Diego County.
From the geographic distribution of vouchers from other herbaria, I almost surely would have found the same results if I had measured the "G. californica" vouchers from San Diego County at any other herbarium housing San Diego County plants. The determination confusion is not due to any problem at a given herbarium; rather, the keys in the floras are poor for separating these two species in Southern California. Using those keys, I never had much confidence in separating these species. I suspect other people trying to determine specimens had the same low confidence in their determinations, too.
My speculation is that the clustering of heads was used to separate vouchers of what were all G. sarothrae specimens into two sets, with the vouchers with less-clustered heads erroneously called G. californica. As discussed elsewhere, the "clustering of heads" is a poorly-defined characteristic, extremely hard to interpret in specimens, and is quite variable even within a single inflorescence from a single plant. If that indeed was the primary characteristic used to determine most vouchers, the analysis here confirms that the clustering of heads is an unreliable separator of the species in southern California.
I thank Jeannie Gregory and Layla Aerne for their considerable hospitality for my visit to the San Diego Natural History Museum on 5 November 2015.
Voucher data for the linked geographic distribution maps provided by the participants of the Consortium of California Herbaria (ucjeps.berkeley.edu/consortium/), retrieved in late September 2015.
In Fig. 2, G. californica is the leftmost pix in the second row, and the rightmost pix in the bottom row.