October 06, 2020

Sphinx chersis: Phenotypic Differences Between Larvae of Eastern and Western Populations



Sphinx chersis (Hübner, 1823), the great ash sphinx, has a rather peculiar geographic distribution throughout the USA and Canada that can be roughly split into at least two distinct populations: 1) an eastern population ranging from northeast USA to southern Canada, and 2) a western population in the southwest USA. (An additional third, southern population exists in Mexico but will not be treated here due to lack of sufficient records.) The species is largely absent in the Pacific Northwest, southeast USA, and most importantly, the central USA, separating the range of the eastern and western populations by considerable distances in all directions. This naturally provokes the question of whether the two populations may actually represent distinct, geographically isolated entities. According to the Sphingidae Taxonomic Inventory, a number of subspecies have historically been described, with the eastern population treated as nominotypical chersis, the western population split into two subspecies, oreodaphne (Edwards, 1874; TL: California) and pallescens (Rothschild & Jordan, 1903; TL: Arizona), and the southern population as mexicanus (Rothschild & Jordan, 1903; TL: Mexico). The two western subspecies, however, have since been resynonymized with the nominotypical subspecies, and thus there is currently no taxonomic distinction between the eastern and western populations.

As a regular identifier of North American sphinx moths on iNaturalist, however, I noticed something unusual about the phenotypes of the mature larvae. Wagner (2005)'s description of the mature larvae is:

a large greenish or pinkish caterpillar with seven long diagonal lines that may be edged with pink. Waxy green over abdominal segments and occasionally T3; lime green below spiracles and over T1-T2 or T3.... Spiracles elongate, central black area ringed with white.... Horn blue or pink.
However, I found that the phenotype of a number of larvae I reviewed, while appearing to be Sphinx chersis, somewhat deviated from his description (if these were actually misidentifications of another similar species I am unaware of, i.e. Ceratomia sonorensis*, well then I guess this study goes down the drain and I sincerely apologize for the time wasted reading this piece). Upon closer inspection, I speculated that these differences may have geographic associations corresponding to the ranges of the eastern and western populations. This prompted me to investigate further by reviewing every larva observation on the site as well as those on BugGuide (while making sure to correct/ignore any that I believed were misidentifications that could confound the study). Through this, it became clear that there were indeed several consistent phenotypic differences between the larvae of the eastern and western populations. Descriptions of the eastern and western populations are discussed as follows:

Eastern Population (Click here for more eastern iNaturalist observations)
The phenotype of the fifth instar larva is perfectly consistent with that described and depicted in Wagner (2005), which is unsurprising given that he likely based his description on larvae taken from eastern populations. With no exceptions, the spiracles have black centers with white rings (any larva observation with orange spiracles on the east is typically a misidentified Ceratomia undulosa). The dorsal abdomen is often waxy green whilst the rest of the body is typically a lime green, as described by Wagner (2005). The oblique lines along the body are usually weakly developed, whitish, and unedged with any other colors in the pure green form. The horn is typically bright blue but occasionally pink. In some cases (14 out of 333 observations reviewed), there are pink forms (and intermediate forms) in which the oblique lines are yellow and heavily edged with pink and the horn is completely pink. In the prepupal stage, the larva usually exhibits no color change except for an occasional light amber hue dorsally (any observations of wandering prepupae that are brownish on the east is likely a misidentified Ceratomia undulosa, or simply a pink form Sphinx chersis).

Fifth Instar

Ontario, CA

Ontario, CA

Ontario, CA (pink form)

New Hamsphire, USA

Wisconsin, USA

Ontario, Canada

Ontario, Canada

Michigan, USA

New Jersey, USA

Western Population (Click here for more western iNaturalist observations)
The larvae of the western population differ in a number of significant ways from those of the eastern population. Perhaps most strikingly, larvae on the west tend to have spiracles with orange centers rather than black ones, especially in the desert southwest region. Black spiracles appear to be rare, except for along the coast (California and Oregon) where they seem to be more common. The dorsal abdomen often lacks the waxy green coloration and sometimes has a shiny sheen. The oblique lines along the body are usually strongly developed and solid white, sometimes thinly edged with dark teal, especially in desert larvae. No pink forms were observed in desert larvae, which suggests they may not be as common (?). There were, however, two pink form larva from Oregon. In the prepupal stage, the larva exhibits a vivid color change to an amber brown or purplish brown dorsally, the latter especially in desert individuals.

Fifth Instar

California, USA

Oregon, USA (pink form)

Arizona, USA

Arizona, USA

New Mexico, USA

Oregon, USA

California, USA

Arizona, USA

Arizona, USA

Colorado, USA

Overall, perhaps the single most striking and consistent difference between the eastern and western populations is that the prepupa becomes deep brown in the west, whereas there is little to no change in the east. The second most striking difference being the possibility of orange spiracles in the west versus consistent black spiracles on the east. While I do not necessarily suggest reinstating the former subspecies of Sphinx chersis (or other taxonomic revisions regarding eastern and western populations) based solely on phenotypic differences of the larvae, I do find these differences quite notable and intriguing. Perhaps the recognition of these larval phenotypic differences may prompt further investigation into the differences between the two populations (adult morphology, molecular sequence data, etc.). In any case, a subtle update of the larval description of the species in the field guides ought to be due, at least.

*The one major caveat of the study is that a possible systemic misidentification of Sphinx chersis larvae with that of another similar species, namely Ceratomia sonorensis in the southwest, could confound the photographic data and invalidate the phenotypic descriptions. I would say, however, that the Ceratomia sonorensis larva depicted in this image link (assuming the identification is accurate) is safely different than any of the Sphinx chersis larvae I reviewed. Ceratomia sonorensis appears to be very similar to it's eastern relative, Ceratomia undulosa and shares the same diagnostic features: unlike Sphinx chersis, there is a patch of black spots on the anal plate and the spiracles have white centers with an orange posterior and anterior edge. Oddly, however, the larvae identidfied as Ceratomia sonorensis on BugGuide appear to be identical to larvae identified as Sphinx chersis both on BugGuide itself and on iNaturalist. I suspect none of them are true Ceratomia sonorensis, and all are misidentifications of Sphinx chersis, except for possibly this one (?). Therefore, I believe most of the larvae identified as Sphinx chersis that I reviewed for this study should be accurate. I did observe quite a large number of misidentifications of other species like Ceratomia undulosa, Paratrea plebeja, Manduca rustica, just about every Smerinthine spp., and even Ceratomia amyntor, etc., but these were quite obvious and thus either corrected or ignored (a funny story is that I had originally attempted to include BAMONA data as well, but upon quick inspection, the rate of misindentifications was so astounding, to say the least, that I decided to abandon the source).

Posted on October 06, 2020 05:36 by alanliang alanliang | 0 comments | Leave a comment

June 07, 2020

Tiger Swallowtail Larva Identification

Tiger Swallowtail Larva Identification

The tiger swallowtails (Papilio [Pterourus] glaucus group) are some of the most charismatic and well-known butterflies in North America. The seven closely related members of the species group include P. rutulus and P. eurymedon (northern Baja California, western US and southwest Canada), P. multicaudata (Mexico, western US, and southwest Canada), P. glaucus (eastern US and southeast Canada), P. appalachiensis (southern Appalachians), P. canadensis (northeast US, Canada, and Alaska), and P. alexiares (northeast Mexico and western Texas). (The latter four species are especially similar and form a species complex, which I will refer to hereafter as the eastern species complex.) Despite the abundance of photographic data for these butterflies, their larvae have long been a challenge to differentiate, even for the most avid of enthusiasts. Larvae are frequently left unidentified or misidentified on iNaturalist and the web and there have not been many distinguishing characters proposed to date. Part of the challenge in identifying character differences is because of the high rate of misidentified larvae on the web, which has likely confounded any serious attempts at revealing a correlation of traits. Here, I have proposed a tentative guide to distinguish tiger swallowtail larvae using character differences for the fifth instar and prepupal stages based on careful surveying of photos from iNaturalist and other open sources, and from personal rearing of P. rutulus and P. glaucus.

Visual Character Differences
Up until the present article, the best known (and probably only) character difference used for differentiating tiger swallowtail larvae involves the inner yellow spot of the eyespot: it is present in all three western species (P. multicaudata, P. rutulus, and P. eurymedon) and absent in the eastern species complex (P. glaucus , P. canadensis , P. appalachiensis, and P. alexiares). The single spot (eastern) vs. double spot (western) character difference is immediately diagnostic on its own and has been recognized for a long time. Though useful to know, most ambiguities between the western species and eastern species complex, however, can already be resolved without visual examination simply by knowing the location of the larva (the range of P. glaucus generally does not overlap that of P. rutulus or P. eurymedon anywhere and only overlaps that of P. multicaudata in Texas and occasionally elsewhere in the Great Plains; P. canadensis meets the three western species only in southern British Columbia and occasionally Alberta). Differentiating between the three western species and members of the eastern species complex themselves, however, is much more challenging, and is discussed in the following sections.

Western Species (P. multicaudata, P. rutulus, & P. eurymedon):
Of the three western species, P. multicaudata is by far the most distinctive. Several character differences readily distinguish it from the other two western species whom it is often sympatric with, the single strongest one being the reduced blue spot in the eyespot pupil that often leaves a ring of ground color within the pupil (2). In addition, there are clear differences in coloration: a brighter ground color in both the fifth instar and prepupal stage (5), a possibility of having a yellow-green eyespot color variant rather than an orange variant (1), and lacking the possibility of having a purple color variant of the blue eyespot pupil (2). Although each of the coloration differences may not be individually diagnostic, when considered collectively, especially with the eyespot shape differences, a reliable diagnosis can be formed. These differences are usually already evident by the fourth instar. Once one has familiarized themselves with these character differences, they may notice that P. multicaudata is frequently misidentified as one of the other two western species and vice versa on the web (the Wikipedia page for P. rutulus for example, blatantly shows photos of P. multicaudata larvae in my opinion). As mentioned earlier, this has likely confounded attempts at identifying diagnostic characters for the species for quite some time. However, an examination of P. multicaudata larva photos with positive adult confirmation or from localities where the two other western species are rare or absent (i.e. Mexico, Texas, parts of New Mexico and Arizona, etc.) clearly shows that the proposed P. multicaudata characters are consistently present. Conversely, the vice versa is also true when examining photos of larvae in localities where P. multicaudata is rare or absent and at least one of the other two western species predominates (i.e. southern California, most of Central Valley) – all larvae consistently lack the proposed characters.

For the other two western species, P. rutulus and P. eurymedon, there are several character differences that can reliably differentiate them from the other tiger species. They are distinguished from P. multicaudata based on the eyespot/pupil shape differences as described earlier (2), and they are also unique among tiger species in having an orange eyespot color variant (1) and a purple pupil color variant (2). Additionally, they are often peppered in minute pale spots, giving them a rougher appearance; these are absent in P. multicaudata, giving it a smoother appearance. However, differentiating between the two species themselves is of great uncertainty and it seems doubtful that a single reliable character difference exists to visually differentiate them. A discussion on BugGuide that is occasionally referenced on other identification sites (erroneously, in my opinion) suggests that P. eurymedon differs in having a "notched" eyespot. However, I am highly dubious of the validity of this character difference, as no other photos of confirmed P. eurymedon larvae support this. Additionally, there was never any positive adult confirmation of the larvae in question, and in fact, I would argue it looks like a very typical P. multicaudata larva.

In my opinion, I suspect P. eurymedon eyespots may be narrower on average and, of the possible color variants, are more typically pure yellow with a purple pupil. P. rutulus on the other hand, has a more even likelihood of exhibiting any combination of eyespot/pupil color variants (1 & 2). In addition, the blue spots on the body might be smaller and darker on average in the fifth instar of P. eurymedon than in P. rutulus. However, these differences may also be dubious and would not be reliable anyways due to the overlap in variability of the examined characters. It should also be noted that the pupil often turns darker and purpler in the prepupal stage which restricts the use of this character difference to only the fifth instar. In terms of the early instar, there may exist a few possible (but probably also unreliable) differences. The shiny, raised bumps (scoli) that stud the thorax, particularly the ones in the eyespots and the horn-like ones behind the head seem on average better developed in P. eurymedon than in P. rutulus in relation to the body size. This is particularly evident in the third and fourth instars as the larvae begin turning green, but the scoli disappear entirely by fifth instar in both species. In addition, the pale saddle and other white markings are more likely to disappear by the fourth instar in P. eurymedon whereas they more likely to persist until the fifth instar in P. rutulus. In the end, however, one will likely need to resort to using non-visual clues for these two species, particularly hostplant (discussed later) and habitat (natural and urban for P. rutulus vs. predominantly natural, montane areas for P. eurymedon).

Eastern Species Complex (P. glaucus , P. canadensis , P. appalachiensis, & Papilio alexiares):
P. glaucus and the three other species in the eastern species complex, P. canadensis, P. appalachiensis, and Papilio alexiares, are unfortunately very difficult, if not impossible to differentiate. Only P. glaucus is included in the above figure as I do not believe characters 1-5 are helpful in differentiating the mature larvae of these four species. Instead, the only difference that I observed between P. glaucus and P. canadensis is that the pale markings in young larvae of P. canadensis are much better developed since birth, which translates to better developed pale stripe/scoli at the rump in the mature (≥L5) larvae (7). (One needs to make certain, however, that the larva they are examining is a fifth instar when using this character difference, and not a green fourth instar, as both species can have well developed pale markings before fifth instar. Mid to late fifth instar is most ideal for diagnosis).
Papilio glaucus vs. Papilio canadensis
Unfortunately, however, the thickness of the pale stripe exhibts clinal variation throughout the intergrade zone in which they are sympatric (northeast US to southern Ontario), and thus the full spectrum of phenotypes can be observed in P. glaucus in this region. The character difference is therefore not very useful in the only area that it is needed. It may be possible however, that many of the intermediate or "canadensis-like" P. glaucus larvae observed in intergrade populations may actually be the larvae of a possible unnamed taxon in the P. glaucus/P. canadensis complex of hybrid origin. The proposed character difference would then partially maintain its integrity in distinguishing P. glaucus from P. canadensis and the unnamed taxon, but the latter two themselves would still be indistinguishable.

For P. appalachiensis, no reference photos could be found so it is assumed to be identical to P. glaucus, whom it is largely sympatric with. The logic for the assumption being that photos that depict P. appalachiensis larvae do exist on the web but the species is so cryptic in regards to P. glaucus that thus far all photos have (unintentionally) been identified as P. glaucus, hence the reason why there is a lack of photos under a P. appalachiensis identification. For Papilio alexiares, only a single image of a prepupal larva could be found and it appears identical to P. glaucus, at least from what can be seen at the angle. Fortunately, however, Papilio alexiares is not sympatric with any of the other members of the eastern species complex, so no character differences are needed to identify it. (It is, however, fully sympatric with P. multicaudata, but the "single spot vs. double spot" and other character differences easily distinguishes them). Like for P. rutulus and P. eurymedon, one may need to resort to using non-visual clues to differentiate members of the eastern species complex, such as date of observation (time of year, uni- vs. bi-/multivoltine), the specific locality, and hostplant (discussed later) in areas of sympatry.

Diagnostic Hostplants
The tiger swallowtail species are unique among Papilio in being polyphagous on a large range of host plant families. Though there is considerable overlap in the host plants utilized between one or more species (ex. Rosaceae, Oleaceae), some families may be locally (or even completely) unique to a species and can thus be diagnostic for identification (6). A more comprehensive (but not exhaustive) list of what I believe to be the more specific preferred hostplants of each species are as follows, with the locally unique/diagnostic hostplants in bold (regional preferences discussed are mostly guesswork):

P. multicaudata - appears to have a preference for Ptelea (Rutaceae) in the southwest US (California to Texas) which is supported by the strong range association between the species and the host in this region; further north where Ptelea is absent (Pacific Northwest, Great Basin, Rocky Mountain states), largely Prunus (Rosaceae) and Fraxinus (Oleaceae); in Mexico, all three commonly utilized, perhaps with preference for Fraxinus (?).
P. rutulus - largely Salix and Populus (Salicaceae) in most of its range, especially inland populations (?). Seems more variable along the coast, often utilizing Prunus, Fraxinus, and Platanus (Platanaceae) as well, especially in urban settings (?).
P. eurymedon - largely Ceanothus and Frangula californica (Rhamnaceae), the former almost exclusively utilized in inland populations (Great Basin, Rocky Mountain states), the latter is a common alternative in California; coastal populations additionally utilize Prunus.
P. glaucus - overall the most polyphagous of all tiger swallowtail species; Liriodendron tulipifera (Magnoliacaeae) and probably related trees (i.e. Magnolia virginiana) preferred in much of its range (except far inland populations where it is absent); Prunus and Fraxinus, and perhaps also Ptelea in some places, also commonly used.
P. canadensis - Populus (and probably also Salix) preferred but may often utilize Prunus and Fraxinus as well.
P. appalachiensis - unknown; only recorded host is Prunus.
P. alexiares - unknown.

Though I am not completely certain on the accuracy of the entire list, what is important are the diagnostic hostplants which can be used for identification. For example, Magnoliaceae and Rhamnaceae are only utilized by P. glaucus and P. eurymedon, respectively, and thus a larva found on such hostplants can often be reliably identified. Other diagnostic hostplants such as Ptelea for P. multicaudata and Salicaceae for P. rutulus can also be for used identification but may require more deduction as they are not necessarily unique to the species. For example, a larva on Ptelea in California is almost certainly a P. multicaudata, while in Texas, it could also be P. glaucus; fortunately, however, the visual character differences (single spot vs. double spot) are already easily sufficient to distinguish the two. A parallel reasoning can also be made for identifying P. rutulus, the predominant utilizer of Salicaceae in much of its range, except where it meets P. canadensis in southwestern Canada; again, however, the visual character differences (single spot vs. double spot) are already sufficient to distinguish these two species.

Posted on June 07, 2020 09:16 by alanliang alanliang | 0 comments | Leave a comment