Plant Species of San Jacinto Mountain:
Comparison of Quercus x morehus, oracle oak to Q. x ganderi, Gander Oak

Fig. 1. Abaxial (bottom) view of oak leaves. Left: Q. x ganderi, Gander oak, a hybrid of Q. agrifolia and Q. kelloggii. Right: Q. x morehus, oracle oak, a hybrid of Q. wislizeni and Q. kelloggii.

The shape of the leaf is sometimes, but not always, different. The Gander oak is often widest near the petiole, or has a nearly uniform width throughout the leaf, whereas the oracle oak is often widest near the tip, as shown by the leaves in these pix. My calculated ratio (see the text) is 0.86 for the Gander leaf, implying it is almost as wide 1/4 of the way up from the petiole as the maximum width of the leaf. My calculated ratio is 0.62 for the oracle leaf, implying the width at 1/4 of the way up is just 62% of the maximum width of the leaf.

The Gander oak generally has cupped leaves, from one of its parents, Q. agrifolia. You can see its turned down edges well in this pix. The oracle oak never has turned down edges.

There is a difference in the number of main secondary veins. Oracle oaks generally have 10 or more main secondary veins, and Gander oaks have 9 or fewer main secondary veins. However, counting the main secondary veins is a bit subtle, and some oracle oaks have fewer than ten main secondary veins. See the text. The vein count in the oracle oak leaf in this photograph is uncertain because the veins at the top of the leaf are too shadowed.

The presence or absence of tufts of hairs in the vein axils (present and circled in both pix) is not diagnostic, since Q. kelloggii leaves sometimes have such tufts of hair, and can pass those genes to either hybrid.

Glabrous leaves are probably always from oracle oak, but both oracle and Gander oaks can have hairy leaves.

Click on the pictures to go to the original iNat observations.

A discussion at iNaturalist for one of my oracle oak observations raised the question as to whether all the plants we have been calling oracle oak at San Jacinto Mountain actually are that determination.

I and others have called our hybrid oaks oracle oaks, Q. x morehus, because we have thousands, and probably tens of thousands, of individuals of each of the two parents, Q. kelloggii and Q. wislizeni, which often grow in close proximity where such hybrids are expected. In addition, there are six vouchers of oracle oak from San Jacinto Mountain, but no vouchers of Gander oak.

In contrast, in the areas where we have found hybrid oaks, there are no Q. agrifolia plants. The nearest specimens of Q. agrifolia are separated in elevation as well as distance. The elevation difference also results in the lower-elevation Q. agrifolia blooming earlier than the higher-elevation Q. kelloggii, making hybrids between those two species nearly impossible.

However, some people at iNaturalist have suggested that at least some of our hybrid oaks are hybrids of Q. agrifolia and Q. kelloggii, Q. x ganderi, Gander oak, based on some of our oracle oaks having hairs on the underneath of the leaves, especially when there are axillary tufts of hairs, which is a hallmark of Q. agrifolia.

So I decided to review the iNat obs of our hybrid oaks, to see if I could resolve this controversy.

I know of two keys to distinguish oracle and Gander oaks, one or both of which were used by some iNat identifiers to determine the San Jacinto hybrid oaks.

According to the key made by Al Keuter, oracle oaks should generally have more than 13 main secondary veins, whereas Gander oaks have fewer than 13 main secondary veins. Al defined a "main secondary vein" as "one that was connected to the primary vein (midvein) on one end and traveled reasonably directly to its terminus at the edge of the lamina".

According to the key made by Fred Roberts, oracle oaks should have glabrescent leaves, whereas Gander oaks should have distinctly-hairy leaves below, either on the entire lower surface of the leaf, or at least in axils of the leaf veins. Gander oaks are found primarily in San Diego County, where Q. kelloggii hybridizes with Q. agrifolia var. oxyadenia, which has densely hairy leaves below.

Fred also mentioned that the shape of the leaves are somewhat different between the two hybrids, with oracle oaks having leaves narrower at the base. To try to capture that difference, I measured the leaf width at its widest point, and also measured the maximum leaf width near the lowest quarter of the leaf, the part nearest the petiole, and took the ratio.

I went through the iNat obs of oracle oaks at San Jacinto Mountain, looking for obs that showed the back of a leaf. I noted how hairy the leaf was, counted the veins, and measure the leaf shape as described above. I was strict about counting only veins that terminated at the edge of the lamina. There are often veins that start looping near the lamina, and don't travel directly to the edge, which I didn't count. If I had done so, my counts would usually be a few veins higher. I also noted whether the leaf was cupped or not, since the leaves of Q. agrifolia are almost always strongly cupped, with the edges of the leaf turned down.

The iNat obs used for my analysis, and my measurements of them, are given in Table 1 at the end of this webpage.

Fig. 2 shows some of the results.

Fig. 2. Plot of Leaf width ratio vs. the # of main secondary veins, for oracle oaks from San Jacinto Mountain. See the text for explanation of these terms.

I found four obs that showed hairy leaves below, two of which were very young leaves, and ten obs that showed glabrous leaves below. There appears to be no difference between the plants with hairy leaves and the plants with glabrous leaves in either the number of secondary veins, or in the leaf width ratio.

None of the obs had cupped leaves, which strongly suggests that these plants did not have Q. agrifolia as a parent.

The number of main secondary veins ranges from 7 to 13 for our plants, with no evidence that there are two populations, one with a higher vein count and one with a lower vein count.

To put our plants into context, I went through the iNat obs of "Gander oaks" in San Diego County, and did the same analysis.

There is a big caveat in the analysis of these "Gander oaks", since some of these might actually be oracle oaks. I.e., San Diego County has both oracle and Gander oaks, making the determination of the iNat obs more difficult. To try to increase the percentage of the obs that were actually Gander oaks, I only analyzed obs that had leaves that were at least moderately hairy below. Note that this may still include some oracle oaks with hairy leaves.

The iNat obs used for my analysis, and my measurements of them, are given in Table 2 at the end of this webpage.

Fig. 3 shows the results. The plot is at the same scale as Fig. 2, so one can flip back and forth between them to see the differences.

Fig. 3. Plot of Leaf width ratio vs. the # of main secondary veins, for Gander oaks from San Diego County. See the text for explanation of these terms.

The most striking difference of the San Diego County "Gander oaks" from the San Jacinto Mountain oracle oaks is that most of the "Gander oaks" have cupped leaves. Eight of the 13 obs plotted in Fig. 3 have obviously-cupped leaves. Only three of the 13 obs have leaves that do not appear to be cupped (one was even anti-cupped!). I could not tell whether the leaves were cupped in two of the 13 obs.

In contrast, not a single one of the oracle oaks from San Jacinto Mountain had cupped leaves.

"Gander oaks" from San Diego County show other significant differences from our oracle oaks at San Jacinto Mountain. They have fewer main secondary veins, 6 to 9 with a single obs with 10 veins, compared to 7 to 13 veins for our oracle oaks. That single obs with 10 veins does not have cupped leaves, calling into question whether it truly was a Gander oak or not.

This difference in the number of main secondary veins is in the same sense as in the key that separates the two hybrids on the number of main secondary veins, but the value that generally separates the two hybrids is 10 veins, not 13 veins. I.e., 9 of the 15 oracle oaks from San Jacinto Mountain have 10 or more secondary veins, whereas only a single "Gander oak" has 10 main secondary veins. It is possible that part of this difference between a dividing line of 10 or 13 veins is due to me not counting some secondary veins. If I didn't count two veins on average, the dividing line would be 12 veins, very close to the 13 in the key.

Gander oaks from San Diego County do tend to have leaves that are of more uniform width, with half of the obs having a leaf width ratio larger than 0.8 (7 out of 12), compared to only 20% of the obs of oracle oaks having the same high ratio (3 of the 15). Since this is only a tendency, it is not very useful as a primary discriminant.

Summary: The hybrid oaks at San Jacinto Mountain are oracle oaks, as expected, not Gander oaks. Their leaves are not cupped; their number of major secondary veins is generally higher than for Gander oaks from San Diego County; and they tend to have leaves narrower at the base.

Oracle oaks at San Jacinto Mountain can have hairy leaves on the underside, and can even have axillary tufts of hairs.

After I did the above analysis, I looked at some leaves of Q. kelloggii at San Jacinto Mountain, and to my surprise found that its leaves can be quite hairy, and can have axillary tufts of hairs (see example one and example two). I then found that I should not have been surprised at this; the Flora of North American treatment for Q. kelloggii says that its leaves have "surfaces abaxially glabrous with small axillary tufts of tomentum to densely pubescent"!

So there is no need to invoke Q. agrifolia as a parent if one find axillary tufts of hairs in the hybrids. That is not a diagnostic trait to separate oracle from Gander oaks.

Speculation. Dave Stith and I have always wondered if oracle oaks appeared differently depending on which parent species was the maternal parent. It is tempting to speculate that the oracle oaks with hairy leaves come from crosses where Q. kelloggii is the maternal parent, and the oracle oaks with glabrous leaves come from crosses where Q. wislizeni is the maternal parent.

Similarly, it is tempting to speculate that Gander oaks with cupped leaves come from crosses where Q. agrifolia is the maternal parent, and Gander oaks with flat leaves come from crosses where Q. wislizeni is the maternal parent.

If anyone knows of experimental crosses where the maternal parent is known, Dave and I would love to know about them.

Table 1. iNat Obs of Oracle Oak Used in the Analysis
Sorted by lower leaf hairs

ObsLeaf cupped?Lower leaf hairsAxil tufts# main secondary veinsratio leaf with 1/4 to 1/2
249390773nodenseno70.86
247863412nodenseyes70.66
110376712nodenseyes80.74
73159865nodenseno90.71
72974743nodenseyes100.73
77476309no(Densely hairy upper)--130.85
292280359nomoderateno130.60
292280358nomoderateyes120.62
100302120nomoderateyes?110.80
72494007nomoderateno110.73
260338955nononeno100.59
170208817nononeno100.78
73161332nononeno110.66
72484835nononeno80.86

Table 2. iNat Obs of "Gander Oak" Used in the Analysis
Sorted by how cupped the leaf is

ObsLeaf cupped?Lower leaf hairsAxil tufts# main secondary veinsratio leaf with 1/4 to 1/2Comments
164566772uncleardense60.50
189927256uncleardenseyes70.89
282004538nmoderateyes100.67No Q. kelloggii anywhere near this plant! This might not be a hybrid oak.
246073469nomoderate80.46
239699735nomoderately70.81Possibly actually an oracle oak.
164566866stronglydenseyes70.85
210683900yesmoderateyes70.50
189340273yesmoderate~yes90.58
246042668yesmoderatemaybe70.72
241330433yesupper leaf hairy80.82
247821680yesmoderate90.83
243501189yesmoderateyes80.83
190444080yesdenseyes90.86


I thank the two iNat people who questioned the determination of our San Jacinto Mountain oracle oaks, @yerbasanta and Fred Roberts, which stimulated this analysis. I now understand the difference in oracle and Gander oaks much better than I did before, as well as learning about the hairs of Q. kelloggii leaves.


Go to:


Copyright © 2025 by Tom Chester.
Commercial rights reserved. Permission is granted to reproduce any or all of this page for individual or non-profit institutional internal use as long as credit is given to me at this source:
http://tchester.org/sj/species/quercus_morehus/analysis.html
Comments and feedback: Tom Chester
Updated 6 July 2025.