[From The Genera of Ichneumonidae, part 1 by Henry Townes (1969. Memoirs of the American Entomological Institute 11: 1- 300.), pp. 7-15.]

HISTORY OF THE HIGHER CLASSIFICATION

The first clear definition of the family Ichneumonidae, and the first arrangement of the genera was by Gravenhorst (1829). His classification is outlined in Ichneumonologia Europaea on a folded sheet at page 64, entitled "conspectus generum et subgenerum". In studying this, one must first realize that Gravenhorst used the term "genus" for a subfamily and "subgenus or family" for a genus. His "genera" were supergeneric groups containing subunits which were treated nomenclaturally as true genera, but designated by Gravenhorst as "families or subgenera". To translate his terminology to modern usage, Gravenhorst's "genera" will be called subfamilies and his "families or subgenera" will be called 'genera. With this translation, he recognized 13 subfamilies, containing 61 genera. In sequence, his subfamilies were 1. Ichneumon; 2. Tryphon; 3. Trogus; 4. Alomya; 5. Cryptus; 6. Pimpla; 7. Metopius; 8. Bassus; 9. Banchus; 10. Ophion; 11. Hellwigia; 12. Acoenites; and 13. Xorides. In this arrangement, the groups with depressed abdomens (numbers 1-8) are placed first and those with compressed abdomens (numbers 9-13) last. Subfamilies 1-8 (abdomens depressed) were subdivided into those with the abdomens petiolate or subsessile (nos. 1-5) and those with the abdomens sessile (nos. 6-8). But Gravenhorst's "Tryphon" (abdomen subsessile or sessile) has a peculiar position between the subfamilies "Ichneumon" and "Trogus". Subfamilies 9-13 (abdomens compressed) also are grouped into those with the abdomens petiolate (nos. 10-11), and those with the abdomens sessile or subsessile (nos. 9, and 12-13). Gravenhorst's general sequence of groups is retained in many later arrangements, which also begin with groups with the abdomens depressed and petiolate, followed by those with the abdomens depressed and sessile, and then by those with the abdomens compressed.

In 1844, Wesmael divided the ichneumonids into 6 groups: "Ichneumones, Crypti, Pimplae, Banchi, and Ophiones", then elaborated the classification of the "Ichneumones". The next general classification was proposed by Holmgren (1855). He recognized five subfamilies: "Ichneumonides, Crypti, Ophionides, Tryphonides, and Pimplariae". These five subfamilies were used by most later authors until about 1945, when a larger number of subfamilies, as were being elaborated by the present author, began to be accepted. Thomson (1873-97) adopted and enlarged upon Holmgren's classification, as did Schmiedeknecht in his Opuscula Ichneumonologica (1902-1933). Thomson and Schmiedeknecht, monographing European species, strongly influenced European usage during their time and up to the present.

In 1868 Foerster published a classification of all genera of the family, which he divided into 36 groups. These groups he accorded family rank. The influence of Foerster's work was boosted when Ashmead, in 1900, published a key to the genera of the family, translating Foerster's keys and adding the subsequently described genera. Foerster's "families" were reduced by Ashmead to tribes and the tribes grouped into the five subfamilies of Holmgren. Foerster's work was far from reliable, and Ashmead, instead of making corrections added his own errors to those of Foerster. Ashmead's compilation was commonly used by workers on the non-European species, especially by those in North America, and partly because of its influence taxonomic chaos flourished.

From about 1895 to about 1925, many new genera were proposed, largely by Cameron, Morley, Schmiedeknecht, Szépligeti, and Viereck. Their efforts contributed many generic names that are at present considered valid, along with enough synonyms and bad taxonomy to confuse themselves and other taxonomists.

In 1920, Cushman and Rohwer published a new classification of the tribes of Ichneumoninae ("Pimplinae") and followed this with revisions of some of the Nearctic genera and species. Their work took into consideration characters of the larvae and of the ovipositor tips. The larvae were not actually discussed, but in personal conversations with Cushman, the present author found that larval morphology and biology strongly influenced his taxonomy. Cushman published additional papers on ichneumonids until 1947. In these some new genera were described and small parts of the family classification revised.

Roman (1903-1943) and Hellén (1914 to the present) used the classification of Thomson and Schmiedeknecht for European species, and contributed improvements. Uchida (1924-1960) wrote on the species of Japan and nearby areas, using Schmiedeknecht's classification with the addition of new genera. Seyrig published on Madagascan ichneumonids in 1932 and 1934, with a posthumous paper on Gelinae in 1952. Heinrich published on the Ichneumoninae of Celebes in 1934 and of Madagascar in 1938. Seyrig and Heinrich, working on exotic species, were forced to re-evaluate the classification, making some significant advances and adding many new genera. Heinrich's work on the Ichneumoninae provided major improvement in the taxonomy of this subfamily. A monograph of the Ethiopian Ichneumoninae by Heinrich, now being published, incorporates further advances.

The present author began to study ichneumonids in 1933. Up until 1952, nearly all of his research was with the Nearctic fauna. It was soon evident that literature on the Nearctic species, except for revisions of a few genera by Cushman and some older papers by Cresson, was so inadequate that it was hardly usable. It was found to be more efficient to ignore the literature, to concentrate study on specimens, and to make identifications by comparisons with types. By 1942, using these methods, a revised classification of the described Nearctic ichneumonids had been assembled. This was published in 1944 and 1945 in the form of a species catalogue. The original catalogue was somewhat revised and published again, in synoptic form, in 1951. In subsequent years these same methods of study and reporting the results were applied to other regions and with each new region investigated, the generic synonymy and general classification were modified according to new findings. A taxonomic arrangement and catalogue of the Indo-Australian ichneumonids was published in 1961, one on the eastern Palearctic ichneumonids in 1965, and one on the Neotropic ichneumonids in 1966.

In the last three catalogues mentioned above, keys to the genera were included. A more extensive work on the genera and higher classification, from a world-wide standpoint, was recently undertaken and its publication is started with the present volume. It is a new classification, with new characters and concepts combined with those that have been appearing in literature since 1944, when the first part of the author's Catalogue and reclassification of the Nearctic Ichneumonidae was published.

NOMENCLATURAL HISTORY

Some of the best known authors in the Ichneumonidae have not been good nomenclaturists, but their prestige has led lesser scholars to adopt and prolong the nomenclatural errors introduced by them. Correcting the accumulated previous mistakes has caused some disturbance in the names, but this has now passed. Some history of the difficulties is recorded here, as an aid to understanding the nomenclature in earlier publications, and as a reminder that vigilance against error is always needed.

Fabricius. The first complication arose with Fabricius, 1804, who pirated several generic names of Hymenoptera from Jurine, 1801, and proposed them as his own new genera, applied to different insects. The name Cryptus was proposed by Jurine for the sawflies commonly known as Arge, but Fabricius appropriated the name for use as an ichneumonid genus. Eventually Cryptus of Fabricius became applied to the genus properly called Itamoplex and was used as the base for the subfamily name "Cryptinae" and the tribal name "Cryptini". Correcting the error originating in Fabricius' action substitutes Itamoplex for Cryptus of authors, Gelinae for "Cryptinae" and Mesostenini for "Cryptini".

Gravenhorst. The next group of errors that persisted originated with Gravenhorst's classification of the Ichneumonidae in 1829. Gravenhorst conscientiously used the generic names proposed by previous authors but was less concerned with applying them to the insects for which they were originally intended. Since Gravenhorst was very influential, most of these errors were continued for about a hundred years. Some of them persist even yet in the papers of certain authors.

Early genotype designations. The next disturbing nomenclatural development was the designation of genotypes by Curtis in 1828-37 and by Westwood in 1840. Their work was good in itself, but it was not heeded by later authors in subsequent usage of the names Pimpla, Mesoleptus, Campoplex, lschnus, and Plectiscus. These names were applied contrary to the genotype designations by Wesmael in 1844 (Ischnus), by Holmgren in 1856 (Mesoleptus), and 1858 (Campoplex), and by Foerster in 1868 (Pimpla and Plectiscus). Authors who followed tended to perpetuate these errors, and usage did not begin to be corrected until Viereck's work on the genotypes in 1914.

Foerster. In 1868 Foerster described 489 new genera of Ichneumonidae, in a paper which had the form of a key to the genera. None of these new genera had species referred to them. In 1871 and 1876 Foerster published revisions of the Microleptinae and Stilpnina, in which he assigned species to the genera of those groups. The history of the remaining genera, without included species, has been controversial and varied, and Foerster's paper has been the source simultaneously for many of the well-known generic names for Holarctic species and for much of the nomenclatural confusion. C.G. Thomson (1873-97) compounded the complication, at first by publishing new generic names in ignorance of Foerster's work, and later by refusing to use Foerster's names. He changed the spelling of such of the Foerster genera as he could recognize, and claimed them as his own new genera. Schmiedeknecht (1902-33) and others continued this confusion by using Foerster's names for some genera and Thomson's names for others, depending on reasons which were undoubtedly at one time logical, but which now seem whimsical.

A large number of Foerster's genera could not be recognized from the descriptions. Studies of Foerster's original specimens have shown that this difficulty stems from erroneous statements of the characters, descriptions of the same genus as new more than once (in different taxonomic groups), and the use of variable or freak characters for defining genera. In some cases his "genera" were assemblages of unrelated forms grouped only on the basis of artificial key characters, or on incorrect or inconsistent observation of the characters. Authors trying to use the Foerster names frequently referred species to them which were far removed from the species Foerster had seen. In some cases these fit the descriptions better than Foerster's own material.

This nomenclatural situation was complex, with no agreed upon guiding principles for simplification. Some advocated that generic names without included species should be considered unavailable until they had received species. Others wished to recognize these names from the date of publication, but with conflicting opinions about the procedure for type selection. The question was referred to the International Commission on Zoological Nomenclature, which in 1912 replied with its Opinion 46, stating that the names should be recognized from the date of publication. For genotypes, it was stated that " ... if (as in Aclastus Foerster, 1868) it is not evident from the original publication of the genus how many or what species are involved, the genus contains all of the species of the world which would come under the generic description as originally published, and the first species published in connection with the genus (as Aclastus rufipes Ashmead, 1902) becomes ipso facto the type."

From several standpoints of nomenclature this Opinion was unsound. It probably was dictated to save names of similar origin already adopted by certain vertebrate zoologists. It gave an authoritative method for dealing with the Foerster names, but for practical reasons it could not easily be put into effect. For the fifty years following 1912 there was nomenclatural turmoil, as the Foerster names were gradually clarified and adopted in accordance with Opinion 46.

A question that hampered the application of Opinion 46 involved the cases where the first species included in one of Foerster's genera obviously did not agree with the original description. It was argued, particularly by Cushman, that a genotype must "come under the generic description as originally published," and if the first species included in a Foerster genus did not run properly to that name in the Foerster key, it could not be the genotype. Cushman published on a number of cases of this sort and proposed corrections. Also, some authors were skeptical when non-European species became genotypes under the provisions of Opinion 46, for Foerster's genera were in fact proposed only for European species.

Now, one hundred years after their proposal, nearly all of the Foerster genera have genotypes in agreement with Opinion 46, and are incorporated into the system of nomenclature. The job was done piecemeal. Woldstedt (1877-81) assigned a few species to Foerster genera, as did Brischke, Schmiedeknecht, Kriechbaumer, Ashmead, Davis, Dalla Torre, Cameron, Roman, Viereck, Uchida, Townes, and a few other authors. The species included in the Foerster genera by Schmiedeknecht and Kriechbaumer have special authenticity. Schmiedeknecht had the original manuscripts of Foerster's generic revision, in which Foerster gave more details than in the publication, including descriptions of the genera and assignment of species to some of them. Schmiedeknecht sometimes quoted directly from these manuscripts. Kriechbaumer, in Munich, had before him the major part of Foerster's collection, with many of the genera labeled by Foerster.

In 1941, when I was completing the Catalogue and reclassification of the Nearctic Ichneumonidae, it was necessary to reach some policy decision on the Foerster genera. Although Opinion 46 of the International Commission, which dealt with the case, seemed somewhat unwise and ambiguous, it could in spite of its faults be made to work, and its application seemed likely to bring ultimate unanimity on the Foerster names.

The argument by Cushman and others that species which did not "come under the generic description as originally published" could not become genotypes, and/or that only European species could become genotypes, had to be considered. After studying some of the actual cases it appeared that whether or not a species fits a description is a matter of opinion, and not always provable one way or the other. About 1921, Foerster's original manuscripts had been sold by Schmiedeknecht to the U. S. National Museum, where they could be studied by the present author. It was evident from the manuscripts that sometimes the very species on which Foerster based a genus would certainly not fit the description. Such a case was reported for Chirotica by Schmiedeknecht (1897. Termesz. Füzetek, 20: 502). Another is Sobas, for which Schmiedeknecht in 1890, had designated Cryptus cinctorius Fabricius as genotype. Cushman argued (1920. Proc. U. S. Natl. Mus. 58: 258) that cinctorius could not be the type of Sobas because it did not fit the original description. Soon after publishing this, Cushman was able to examine Foerster's original manuscript and found that Foerster erected Sobas precisely for the species cinctorius. Cushman later pointed this out (1927. Proc. U. S. Natl Mus. 72 (13): 6).

Foerster's work contained enough error that lack of agreement with one of his descriptions could not be proof that a species was not one on which he based a new genus. Also, since whether or not a particular species fits a description is always a matter of opinion, and since someone's opinion has to be accepted, it is most convenient to accept the opinion of the author who first referred the species to one of Foerster's genera. If that author believed that a species fit well enough to refer it to the genus, nothing important would be gained, and stability would be lost, by discarding the first author's opinion in favor of the opinion of some later one.

In the author's Catalogue and reclassification of the Nearctic lchneumonidae (1944-1945) the first species (or one of the first group of species) definitely referred to a Foerster genus was accepted as genotype. This policy has since been generally accepted and now seems firmly established. If it had not been adopted, stability for the Foerster names would not yet be in sight.

Though agreement on policy was reached more than 20 years ago, the factual foundation for execution of such a policy (or any other policy) was not adequate. Some careful and exact scholar with an understanding of nomenclature, knowledge of the Ichneumonidae, and familiarity with the literature on Ichneumonidae had to determine which species qualified as genotypes for the hundreds of Foerster genera. Moreover, nearly a century of using the Foerster keys had not brought forth any species to fit several score of them. Those species-less generic names were still derelicts, and a threat to stability because they could unexpectedly upset the names of subsequently proposed genera. The bibliographic work had to be supplemented by a search of the remnants of Foerster's collections in the museums in Munich, Berlin, and Vienna, to see if there were specimens of the genera labeled by Foerster himself. The person who elected to do all of this was Dr. J.F. Perkins of London. His results were published in 1962. Foerster's genera have been a monumental problem in nomenclature, perhaps never again to be equalled. Dr. Perkins' answer to the problem is as monumental as the problem itself, a model of accuracy, thoroughness, and the precise application of nomenclatural principles.

After search of the literature and of Foerster's collections, twenty-five genera were still without genotypes in Perkins' publication. Subsequent work, however, has supplied genotypes for nearly all of these twenty-five.

Thomson. In 1873 and 1874, C.G. Thomson published his first two papers on the Swedish Ichneumonidae (Opuscula Entomologica, fascicles 5 & 6). These papers treated the “Crypti" [Mesostenini and Echthrini], with the proposal of many new genera. Thomson was originally ignorant of the Foerster genera published in 1868. He became aware of them prior to publication of fascicle 6 and in fascicle 6 (pp. 590-591) discussed some of Foerster's new genera. He did not, however, withdraw his new names in favor of the prior ones of Foerster. Also in subsequent papers he refused to adopt the Foerster names, except in the Microleptinae and Stilpnina where Foerster himself had included species. Where he knew the application of Foerster generic names and considered the genera zoologically valid, he proposed his own new names, these being usually emended versions of the names used by Foerster. The superior quality of Thomson's taxonomic work and the inconvenience of discarding the generic names proposed by him tended to perpetuate the use of his generic names even by authors who conceded that Foerster's names should be recognized as prior.

Though there developed agreement that the Foerster names had priority, there was confusion about how to treat the Thomson variants of them. Some authors believed they constituted proposals of separate new genera, concluding that Thomson rejected Foerster's publication as not worthy of recognition, and proposed new names of his own, using Foerster's work as a mine for the names as he might use an unpublished manuscript. With this view, the Thomson names could have genotypes separate from those of the Foerster names of similar (but different) spelling. This is the viewpoint adopted by Viereck in his work on the ichneumonid genotypes (1914), although in this he was not entirely consistent. Cushman also had this opinion, and after conversations with Cushman on the subbject, and examination of Thomson's work, the present author adopted the viewpoint of Cushman and Viereck. Thus the Thomson names were treated as separate generic names, occasionally with separate genotypes, in papers by the present author prior to 1961. While this viewpoint was defensible, probably correct, and had the force of a tradition started by Viereck, it came under attack from Walkley in 1958 and from Perkins in 1962.

The Walkley and Perkins opinion was that Thomson considered the Foerster names as nomenclaturally available but emended them slightly to suit his own views on the spelling of scientific names. On review of the evidence it seems that Walkley and Perkins do not have a provable case, but neither is it "provable" that the view of Viereck and Cushman is the only logical one. Thomson, in fact. was ambiguous, and his work also varied in its relation to the Foerster genera from one year to another. A conclusion about how to treat his names would be influenced by what parts of Thomson's work were emphasized and what parts passed over, and a decision about which course to follow would be partly arbitrary. In 1961 and afterward, the present author changed to the viewpoint of Walkley and Perkins (that the Thomson names are emendations and therefore must have the same genotypes). It was realized that, though the Viereck and Cushman treatment could be supported by some of the evidence, it frequently led to awkward nomenclatural situations where two very similar names were treated as separate genera. So, in this case where theoretical considerations are not entirely decisive, practicality is permitted to rule.

Viereck. In 1914 Viereck published a list of the genera of Ichneumonidae, with their genotypes (and a supplement in 1921). Many new genotypes were designated. Until recently, this list served as a basic reference for the generic names. Viereck's bibliographic data were not adequate, however, so that certain of the genotypes are incorrect, particularly in the cases of Foerster genera. Viereck used Dalla Torre's Catalogus Hymenopterorum (1901-02) as the source for much of his bibliographic data rather than personally searching the literature, and some of the generic synonymy and bibliographic references in the present author's earlier catalogues are erroneous because of too much reliance on Viereck. Viereck's work in regard to the Foerster names was thoroughly reviewed and corrected by Perkins (1962), and for the present monograph, this author has personally verified the original descriptions and type designations of all the genera.

Viereck's genotype list has two noteworthy peculiarities: When a genotype was selected, the species with page priority was favored. This first species was sometimes not typical of the genus or was rare in collections, and with its establishment as genotype there were sometimes unfortunate shifts of names or other inconveniences. The second peculiarity is that when a genus was monobasic and the genotype also designated, Viereck gave the genotype fixation only as "monobasic". He believed that in a monobasic genus a type designation was superfluous. He was, of course, correct, but authors who emphasize designations of genotypes cannot always find in his list the data that they may wish to quote.

References