The Gall Wasp Genus Cynips: A Study in the Origin of Species/Phylogenetic History
PHYLOGENETIC HISTORY
Analyses of the phylogenetic history of any group of organisms and biologic interpretations of taxonomic data depend for their validity upon the soundness of the available classifications of the group. If the catalogs are poorly made,one may draw no conclusions or, what is worse, draw conclusions as fantastic as the hobgoblins of primitive imagination. But if the taxonomic arrangement brings together species of common ancestry and accurately portrays the varying degrees of relationship between those species, a classification becomes one of the most powerful tools available for the evolutionary interpretation of biologic phenomena. It becomes a code by which one may translate the biologic and distributional data into the story of the origin and paths of dispersion and the order of development of each species and of each biologic characteristic of a group, from its primitive beginnings and thru the several stages by which it evolved the peculiar phenomena which we find today.
Phylogenetic interpretations of the genus Cynips have heretofore been impossible because cynipid genera, in common with the genera of many other insects, have been established for the most part upon “diagnostic” characters of insect morphology. These have been drawn from the toothed tarsal claw, the dorsally produced and naked abdomen, and the hairy thorax of the agamic form of the species folii, the genotype of the group (see Mayr 1870-1905, Dalla Torre and Kieffer 1910, Beutenmüller 1911, and Weld 1922-26, where the names Dryophanta or Diplolepis are used instead of Cynips). The insects included in the genus thus defined differed in many points of structure which, however, were consistently ignored. The genus included both black oak and white oak species, species that live in galls on flowers, leaves, stems, and roots, and galls of every conceivable type of structure (e.g., see plates 12 to 17 in Beutenmüller 1911). There were species with divergent types of life histories. There were species that we shall ultimately have to assign to 8 or 10 distinct and largely unrelated genera. The extent to which our own interpretation differs from previous treatments becomes evident in the following table.
Author | Species assigned to present genus |
Accepted in present revision |
Mayr 1870-1905 | 17 | 88 per cent |
Dalla Torre 1893 | 42 | 48 per cent |
D. T. and K. 1902 | 50 | 46 per cent |
D. T. and K. 1910 | 68 | 41 per cent |
Beutenmüller 1911 | 43 | 32.5 per cent |
Weld 1922-26 | 46 | 37 per cent |
Our present delimitation of Cynips is based upon no single character—indeed, we know of no character by which the group may be separated from other groups of oak-inhabiting Cynipidae—but upon a striking, even if not invariable, correlation of insect structures, gall characters, host relationships, life histories, and distributional data that are coordinated for the 93 species which we now bring together.
In nearly every one of the species the agamic form has the thorax hairy, altho there are few hairs on the very small, nearly wingless insects which are the most northern varieties of many of the European and American species in Cynips. In every species without exception the hypopygial spine is broadest posteriorly, and the structure terminates in a well-developed tuft of hairs. A similar spine is found, however, in one or two related genera. In 90 of the species the tarsal claw is rather strongly toothed, but in 3 species of a single stock (Cynips mellea) the claw is so weakly toothed that it is nearly simple. In nearly all of the species the agamic gall is fundamentally spherical and monothalamous, rarely with any remarkable development of the epidermal layer of tissues; but in 8 species of one stock (Cynips pezomachoides) the gall is usually polythalamous. In every one of the 93 species the gall originates from leaf veins, usually on the under surface of the leaf, but in two species (C. multipunctata group) the gall may occur on the leaf petioles and young stems as well as on the leaf. In every one of the 11 species for which the alternating generations are known, the bisexual gall is a thin-walled, seed-like or bladdery, hollow cell located within the newly opened buds of the oaks. All of the 93 species occur on white oaks (Leucobalanus). In nearly every case the agamic generation begins development in early summer, matures by early fall, lies as an adult within the gall for several weeks or months, and finally emerges in the winter. Only the 7 species of the fulvicollis stock modify this procedure by delaying part of the emergence until a second or later winter. Some of the species of the mellea stock also depart from the typical life history by emerging in the early spring instead of the winter season. In distribution the insects show their affinities by occurring in adjacent areas, so the 93 species may be put into 28 groups and these in turn into 6 subgenera which constitute two main groups, each of which is a closely compacted unit in its geographic distribution.
Thus, while the genus cannot be based on any single diagnostic character, the remarkable coördination of so many characters in so many species and the absence of anything approaching this combination of characters among any of the other members of the family testify to the phylogenetic unity of Cynips. There seems no reason for believing that such a body of coördinated characters could have arisen independently in more than one time and place. In a single period, in a limited area, there must have existed a population from which all of the present-day species, with their varying grades of relationships, have developed. This history of the expansion of a single genetic stock into 93 distinct populations, by processes of mutation, isolation, and on occasion subsequent recombination into hybrid populations, is the story we have been unfolding in this study, and which we are now ready to fit into the geologic time and the geographic areas in which speciation probably proceeded in the genus.
Cynips is in every respect a highly specialized genus of the oak-inhabiting tribe Cynipini of the family Cynipidae. The fossil record of the Cynipidae is meager and without significance except to prove that the family was in existence in the Oligocene and Miocene (Kinsey 1919; Cockerell 1921). There seems no reason for believing that the Cynipidae have ever been associated with any plants except the Angiospermae on which the family occurs today. The diversity of the present-day genera of the primitive gall makers of the Aulacini indicates that that tribe must have had a long history antecedent to the origin of the Cynipini. The family could not have originated before the rise of the flowering plants in the late Cretaceous, and it was probably much later before the Cynipini developed such specialized genera as Cynips and Disholcaspis.
Our attempts to fathom the history of any of these higher genera must proceed on the assumption that all species of these groups have from the first been associated with oak, and our analysis of the cynipid history must do no violence to the known history of the sources and development of Quercus. To this end, the accompanying summary of the paleontological record of American oak will serve as reference for some of the considerations that follow.
(Compiled from Trelease, 1924, and Berry 1923)
Area | Cretaceous | Eocene | Miocene | Pliocene | Pleistocene | Living oak flora | |
Northeastern Siberia | x | ||||||
Northwestern North America— | |||||||
British Columbia | x | x | Southern | ||||
Alaska | x | x | |||||
Pacific Coast States— | |||||||
California | x | x | x | x | |||
Oregon | x | x | x | ||||
Washington | x | x | x | ||||
Desert States— | |||||||
Idaho | x | x | |||||
Nevada | x | x | x | Southern | |||
Mountain States— | |||||||
New Mexico | x | x | x | ||||
Colorado | x | x | x | x | |||
Wyoming | x | x | x | Northeastern | |||
Montana | x | x | x | ||||
Great Plains Area— | |||||||
Kansas | x | Eastern | |||||
Nebraska | x | Eastern | |||||
North Dakota | x | Southeastern | |||||
Middle West— | |||||||
Illinois | x | x | |||||
Southern States— | |||||||
Kentucky | x | x | |||||
Alabama | x | x | x | ||||
Tennessee | x | x | |||||
Mississippi | x | x | |||||
Virginia | x | x | x | x | |||
North Carolina | x | x | x | ||||
South Carolina | x | x | |||||
Florida | x | - | |||||
North Atlantic States— | |||||||
District of Columbia | x | x | |||||
Maryland | x | x | x | x | x | ||
West Virginia | x | x | |||||
Pennsylvania | x | x | |||||
New Jersey | x | x | x | x | |||
New York | x | x | |||||
New England— | |||||||
Vermont | x | x | |||||
Greenland | x | x | |||||
Iceland | x | ||||||
Spitzbergen | x | ||||||
South America | x | Mts. of Colombia |
The ancient history of Cynips is first of all to be read in the characters and distribution of the six existent subgenera. The range of each subgenus is shown on the accompanying map (fig. 7), and the data are further summarized:
Subgeneric Ranges of Cynips
Subgenus | Species | Range |
Cynips | 11 | Europe, Mediterranean Africa and Asia, possibly elsewhere in Asia. |
Antron | 12 | U. S. Pacific Coast, California-Oregon, undescribed
species in southern Arizona. |
Besbicus | 8 | U. S. Pacific Coast, California-British Columbia. |
Philonix | 8 | U. S., Arizona-Atlantic Coast. |
Atrusca | 12 | Mexico; U. S., Arizona-Atlantic Coast. |
Acraspis | 42 | Mexico; U. S., Arizona-Atlantic Coast. |
These subgenera clearly represent two groups, which may be distinguished as follows:
Cynips-Antron-Besbicus
Philonix-Atrusca-Acraspis
FIG. 7. KNOWN DISTRIBUTION, SUBGENERA OF CYNIPS
Base from Goode series of Base Maps, by permission University of Chicago Press.
The primary subdivision of the genus into one group that is all but exclusively Eurasian and Pacific American, and into a second group that is confined to North America east of the Sierra Nevada, should have occurred near the center of the origin of the genus. We may hypothesize this center in the southwestern United States or in adjacent areas of northern Mexico. From here the first subgeneric group could have moved westward to the Pacific Coast where Antron and Besbicus were isolated, and by way of Alaska and Siberia into Asia and Europe where the subgenus Cynips developed. The second group, differentiated into the subgenera Philonix, Atrusca, and Acraspis, could have spread to the north in the Rocky Mountain area, to the south in Mexico, and to the east toward the Atlantic Coast. On the other hand, if the genus were assumed to have arisen in Europe or Asia, it is difficult to understand why the Eurasian affinities should have been maintained onto the Pacific Coast of North America, and a primary subdivision of the genus not effected until the group reached the Southern Rockies. However, we are completely ignorant of the present-day occurrence of the genus in most of Asia (on pp. 447 to 453 we show that the species reported from Japan do not belong to the genus), and further discoveries may throw light on Asiatic beginnings of the group; but from the available data we would presume the southwestern American origin for the genus and its initial differentiation in that area.
That the point of origin was not north of central Arizona or New Mexico is indicated by the occurrence of numerous short-winged species both north and east of the region, while not a single short-winged species of the group is to be found in the southern halves of those states. We have already shown that the short-winged species represent more specialized developments of the long-winged stocks. One short-winged and one long-winged species of Cynips is known from central Mexico, but our knowledge of the gall makers of that country is still insufficient to make it certain that our genus did not originate somewhere in northern Mexico.
It is interesting to find that Trelease (1924:34), considering the origin of the American oaks, states that “their primary center of distribution appears to have been what is now Arizona.” Trelease's conclusion is based in part upon the opinion that the widely distributed Cretaceous oaks have left no descendants in present-day groups, and that the existent white oaks of Europe and Asia, and both the white and black oaks of America have developed from the type represented by the European, Tertiary Quercus Palaeo-Ilex. This type is presumed to have reached America at some time previous to the Cenozoic. Berry's criticism (1923:139) of this conclusion seems to be based upon a mis-interpretation of Trelease's involved presentation of the argument, but the matter needs more data than seem to have been derived from the Cretaceous fossils that are often dubiously referred to Quercus.
The close affinities of the Eurasian and Pacific American subgenera of Cynips, and the more unique nature of the eastern American subgenera suggests that the migration between Eurasia and North America was by way of the Alaskan-Siberian land bridge, rather than by way of former land connections between Labrador, Greenland, Iceland, and northern Europe.[1] The migration of the first group of Cynips
FIG. 8. PHYLOGENETIC ORIGINS, SUBGENUS CYNIPS
from its center of origin in the Southwest, to our Pacific Coast, and finally across Alaska into Siberia, must have occurred before the Great Basin became arid, and while the Alaskan-Siberian land connections were still enjoying a climate mild enough to have supported an oak forest. From the preceding table it will be seen that fossil oaks are known to have occurred in both Siberia and Alaska as late as the Eocene, and the land connections between the two continents were continuous thru the late Miocene and intermittently existent at later periods. Berry, however, expresses it as his opinion (1923:140) that “it is very doubtful if there was any possibility of an interchange of species of oak between the Old and New Worlds after late Eocene or Oligocene times,” and this may be the latest origin we may presume for the genus Cynips.
On the other hand, it is difficult to understand why both the primary subdivisions of the genus did not spread both to the east and to the west of their Southwestern center of origin if they migrated very much before the Great Basin became so arid as to interpose an effective barrier to further interchange of eastern and western species. It is understandable that each group might have begun its migration in a particular direction, but our knowledge of the present-day distribution of organisms would lead us to expect that a group should, in sufficient time, radiate in every direction in which there are no recognizable barriers to migration. But the rising mountains of the Pacific Coast probably did not bring about the development of the Great Basin deserts until the Miocene. If the Eurasian-Pacific-American branch of the genus crossed on to the Coast in the Miocene, its failure to radiate in other directions might be explained as due to the subsequent development of aridity in the Great Basin before the group had time to move back across that area. But if the group must be taken out of the Southwest and across Alaska at an earlier date, as Berry's statement would require, it is more difficult to understand why the Pacific Coast subgenera did not have an opportunity, before the Great Basin became arid, to spread back into the more eastern United States.
In the southernmost mountains of Arizona there is a living variety of the Pacific Coast species Cynips (Antron) guadaloupensis. This is the only representative of that subgenus known from east of the Sierras, but it is matched by a few cases in other cynipid genera and by a few Pacific slope trees, reptiles, and other organisms that have stray relatives in southern Arizona. These strays are, however, such close relatives of existent Californian species that they are probably to be interpreted as more recent arrivals in Arizona rather than remnants of the primitive stock before it moved westward into California. During the Quaternary the Great Basin had a more moist climate than it had had since the Miocene, and the area just north of the Gulf of California then may have supported enough oak to have allowed the extension of some species of the typically Californian fauna.
Within California the paths of migration of the several species of Antron and Besbicus probably began at some point in the eastern part of the state and extended north and south and about the Great Valley. The Valley was not completely cut off from the sea until the Pliocene and Pleistocene. The
present-day range of many plants and animals of the foothills of the mountains rimming the Great Valley may be due not only to differences in topography and climate and vegetation at different elevations, but to the influence of the more ancient distribution as well. Altho many Cynipidae range all the way from Bakersfield in the southern end, to Shasta Springs at the extreme northern end of the Valley, and altho many of the species of the foothills and even higher elevations of the Sierras find no barrier in the latitude of San Francisco, many of the insects of the Coast Ranges have northern varieties that reach their southern limits and southern varieties that reach their northern limits near San Francisco Bay. If the migration in such cases had been wholly from the north or wholly from the south, it is not easy to understand why the break should occur near San Francisco Bay; but the situation is explainable if it is presumed that both stocks originated from the eastern Sierras, that the northern variety reached San Francisco Bay from the north and the southern variety from the south, and that the Bay never was crossed until a geologically recent day.
The Eastern American group of subgenera of Cynips was differentiated in the late Miocene or early Pliocene, as the following considerations may show.
None of these subgenera are represented west of the Great Basin today, and it is probable that they were prevented from reaching the Sierras by the development of the Great Basin deserts in the Miocene.
That the subgenera were distinct and most if not all of the present-day species differentiated before the end of the Pliocene is attested by the fact that all but one of the species found east of the Great Plains is represented by close relatives, either very closely related species or varieties of the same species, in the Rocky Mountain area. The following table summarizes the situation.
Eastern Species | Rocky Mountain Relatives |
C. fulvicollis | C. plumbea |
C. centricola | C. dugèsi and C. bella |
C. pezomachoides | C. pezomachoides, 1 variety |
C. gemmula | Not known |
C. hirta | C. hirta, 2 varieties |
C. villosa | C. villosa, 5 varieties |
C. mellea | C. mellea, 1 variety; and C. arida |
This eastern extension of the Rocky Mountain fauna must have occurred before the Rockies reached their heights and thus caused the aridity of the Great Plains in the Pliocene. Today, between the easternmost oaks in Colorado and the westernmost extension of oaks in Kansas there are three or four hundred miles of Plains that are now barren of oak. Between the Rocky Mountain oak of northern Colorado and the westernmost extension of the eastern Quercus macrocarpa in northwestern Nebraska is a stretch of fully two hundred miles. In the Black Hills of the South Dakota-Wyoming boundary, this same eastern oak comes into contact with the Rocky Mountain Quercus Gambelii, but this could at most
FIGS. 11-12. PHYLOGENETIC ORIGINS IN PHILONIX AND ATRUSCA supply a present-day means of eastern migration for only those species of Cynips that occur on the Gambelii-macrocarpa oaks. There are two such species, Cynips villosa and C. hirta. Villosa may, for all that is now apparent, have crossed from the more northern Rockies. On the other hand, the more southern concentration of the varieties of C. hirta, in part upon the chestnut oaks, suggests that this species came eastward by the Colorado-Missouri route which, we shall show in a moment, was followed by the remaining species of the genus. Any oak-inhabiting cynipid that crossed in this part of the Plains must have done so before the extermination of the oak flora in those areas in the Pliocene.
It is possible that during the southernmost extensions of the glaciers of the Pleistocene some increase in moisture allowed oak to return to some southern parts of the Great Plains. It is certain, however, that the several stocks of Cynips had come east before then, because in the northern Middle West there are several species which, as we have shown (page 59), seem to have had a hybrid origin in the Pleistocene. If the southern extension of the glaciers at that time crowded northern varieties into the ranges of southern varieties of the same species, with consequent hybridization of the close relatives, it follows that northern and southern varieties were already differentiated in the eastern United States.
Altho eastern and western species of oak make rare contacts in the Texas Panhandle and in northeastern New Mexico, the affinities of all the Cynipidae of Texas east of the Pecos River are clearly with those of the eastern United States, while all the Cynipidae of West Texas are of more direct origin from the Arizona-New Mexico stocks. The geologic record indicates that the desert boundary between West Texas and the more eastern part of the state is more ancient than the genus Quercus. One may conclude that the cynipid fauna of eastern and central Texas has been derived from the north and east and not directly from the Southwest.
The present-day concentration of the species of Acraspis (fig. 49) indicates that the eastern migration did not occur very far north of Texas. It probably occurred in Kansas and Missouri, or not far north or south of the boundaries of those states.
FIG. 13. PHYLOGENETIC ORIGINS IN ACRASPIS The chief path of migration of Cynips moving eastward from Missouri probably bent southward about the southern Appalachians and finally up the Atlantic Coastal Plain. The present occurrence of so-called Coastal Plain Cynipidae, or close relatives of Coastal Plain species, across the state of Missouri, thru adjacent parts of the Mississippi Valley, and along the course of the Tennessee River, as well as out on the Atlantic Coastal Plain itself, probably records the path of migration from the Rockies to the easternmost limits of the United States. In the systematic portion of this study data bearing on this point are recorded and mapped under:
Cynips centricola centricola |
C. pezomachoides pezomachoides |
C. pezomachoides derivatus |
C. gemmula fuscata |
C. mellea carolina |
The restriction of the short-winged species of each of the stocks of the eastern subgenera to those points which are furthest along this supposed route out of the Southwest, is some further verification of the route. Spreading to the north and to the south of this main path, the ancestral stocks gave rise, by mutation and isolation and, in the few cases noted, by subsequent recombination of characters in hybridization, to the numerous populations which are the species or varieties today. This history is summarized in the phylogenetic maps.
Many will find the present-day species of our genus not easy of identification. There are some who are inclined to believe that species are, after all, but human concepts instead of realities in nature. Some will consider that individuals are so variable and interbreeding Mendelian races so abundant in nature that taxonomic classifications can be nothing but contrivances without biologic significance. And yet, when the Colorado of the West first cut into the Colorado Plateau the specific stocks of Cynips were in existence, and thruout the years that the Canyon has been cutting, even down onto the present, these complexes which we call species have maintained their identity. While the eternal hills have come and gone, these instable protoplasmic, entities have maintained their stability. A stability like that of a stream, with materials always contributing from many sources, with endlessly Phylogenetic History
Epoch | Time estimate, millions of years |
Geologic events | History of Cynips |
Upper Cretaceous | 80-60 | Rise of flowering plants Origin of oak First rise of Rockies and Sierras |
Ancestral Cynipidae |
Eocene | 60-35 | Alaskan-Siberian connection maintained, climate mild, oak in the area | Differentiation of main groups in the family |
Oligocene | 35-20 | Alaskan-Siberian connection maintained, climate mild, oak in the area Great Plains forming |
Cynips and other modern genera in existence Eastern and western subgenera separated |
Miocene | 20-7 | Alaskan-Siberian connection continuous, climate more severe Rockies elevating, deserts forming in the area Finally Alaskan land connections broken |
Subgenus Cynips moves into Eurasia Antron and Besbicus species differentiate on Pacific Coast Acraspis, Atrusca and Philonix specific stocks arise in southern Rockies |
Pliocene | 7-1 | Great Valley of Calif. in part an inland sea Rockies elevated and Great Plains become semi-arid Sierras elevated and Great Basin becomes more arid |
Specific stocks of eastern subgenera move across Great Plains Modern species and varieties differentiated in all groups |
Pleistocene | 1 million to 20,000 years | Grand Canyon begun Extensive glaciation Cold climate south to Ohio River Moist climate in parts of Great Basin and north of Gulf of California |
Northern varieties pushed south to hybridize with southern varieties Some Calif, species migrate back to so. Arizona |
Tho the taxonomist's specimens seem insignificant, the data tedious, and the dry-rot of the technic unendurable, the pinned specimens in the box are a bit of the forests and the hills and the days where the drama of wasp life is unfolded, an evidence of ancient origins, mutating genes, and inexorable marches of evolution, an epitome of the enduring and everlastingly changing entities which are species.
- ↑ Thruout this part of this study I have had the criticism of Dr. C. A. Malott and Dr. J. W. Beede of the Geology Department of Indiana University. Geologic data pertaining to this section are summarized in such texts as Miller's Introduction to Historical Geology (1916) and Schuchert's Historical Geology (1924). Berry (1923) and Trelease (1924) summarize the paleontological record of Quercus.