part of my program of revision of gall wasp genera. All of the generic terms used in this paper define distinct, phylogenetic groups with the exception of the term Andricus. A good genus Andricus can be separated some day, building about the type species, but until then the term covers a waste basket. I do not mean to imply the close relationships of the species here placed in Andricus, but rather to indicate that their generic positions cannot yet be defined. I employ Cynips to indicate. a good genus, of which Cynips folii Linnæus is the type (designated by Westwood, 1840, Synop. Gen. Brit. Ins., p. 56, and reaffirmed by Rohwer and Fagan, 1917, Proc. U.S. Nat. Mus., LIII, p. 364). The majority of the species placed by both European and American authors in Cynips do not belong to that genus, and some of those placed in Dryophanta Förster (isogenotypic with true Cynips) and Diplolepis Dalla Torre and Kieffer (not Diplolepis Geoffroy!) belong to true Cynips. Cynips is here used in a restricted sense which I shall more fully explain later. I am not using the name Callirhytis because its type is unrecognizable. In the original establishment of the genus Förster included his Callirhytis hartigi and designated it as type. Hartigi was poorly described, without a host record, and without a locality record, altho probably the insect came from Germany. As far as I know, the insect has never again been recognized. Later authors have more or less generally adopted characters to restrict the term, particularly that of a simple tarsal claw. However, it is not known whether hartigi had simple or toothed claws, and until we again recognize that species the later restrictions on the term Callirhytis cannot have any standing. Certain it is that the group as generally defined today in no sense defines a phylogenetic unit, and is meaningless. I prefer to reduce these meaningless terms, and shall use only Andricus in that way.

Some years ago I heard Dr. W. M. Wheeler remark that what we needed in taxonomy was not longer, more detailed descriptions, but pointed comparisons of related forms. These comparisons are possible only when much material is available. As far as possible I have used the method, and I think with highly satisfactory results.

I have no interest in taxonomy per se, for it is not a science concerned with questions of cause and effect. But until the foundation of cynipid taxonomy is developed we shall not be

able to build the structure of cynipid biology. The students of cynipid taxonomy must be the ultimate critics of the biologic hypotheses we build, and for that reason I have included a considerable body of biologic data with the taxonomic record of each insect. These data are brought together in the index. After the taxonomists have criticized, I shall draw these data together into a more general form.

The material on which these studies are based was mostly collected during 1919 and 1920 while I held a Sheldon Travelling Fellowship of Harvard University. Fully half of the new Cynipidæ I obtained at that time are not yet described, and the volume of biologic data is not yet touched. This is the second paper I have based largely on this material, the first having been published recently (1922, Bull. Amer. Mus. Nat. Hist., XLVI, pp. 279-295), and others should follow soon. Great credit is due the Harvard University authorities, especially Dr. William Morton Wheeler, who made the trip possible.

Eleven of the new varieties are described from material which was collected and bred by Dr. J. T. Patterson, of the Zoology Department of the University of Texas. Dr. Patterson is undertaking probably the most extensive work which has yet been done on cynipid life histories and genetics. I greatly appreciate the opportunity to describe his material, most of which I have not yet been able to study. Mr. C. W. Johnson, of the Boston Society of Natural History, Mr. William Beutenmuller, and Dr. A. L. Melander collected types, each of one of the new varieties.

A great number of friends have contributed largely to my work. Mrs. Kinsey has generously contributed encouragement, time, and skillful criticism. I have secured gifts or loans of material, particularly types, and other favors from the Museum of Comparative Zoology, the American Museum of Natural History, Mr. S. A. Rohwer and Mr. L. H. Weld at the U.S. National Museum, Dr. Isabel McCracken at Stanford University, Pomona College, Professor A. Trotter of Portici, Italy, and a host of friends who are not biologists. Mr. Frank A. Leach, of Diablo, Calif., is sending me material. Dr. F. E. Lutz and Mr. Andrew J. Mutchler, of the American Museum, have especially aided the work. The authorities of Indiana University, particularly the members

of the Zoology Department, have been liberal in their encouragement and material help. Mr. S. B. Parish, then of San Bernardino, Calif., introduced me to the Pacific Coast oaks. Dr. Forrest Shreve helped very considerably in my travel in Arizona. Mr. W. H. Vance and Mrs. I. T. Wilson have capably worked in mounting my material for study. For all of this help I am greatly indebted. Most of this work would have been impossible without such coöperation. I have not had access to the paratypes in the private collections of two of the cynipid workers, and I feel that if any injustices have been done their species some of the blame is due them.

The holotypes of all these new Cynipidæ, with one exception, are in The American Museum of Natural History. Paratypes have been distributed as widely as possible among the museums, and I shall be glad to lend or give paratypes or other material, as long as it is sufficient, to students of these insects. Any material of these species labelled cotypes should be considered paratypes.

Andricus californicus (Bassett)

FEMALE. Brownish rufous, head and thorax hairy; median groove about lacking; areolet large; first abscissa of the radius sharply angulate; length 3.0-5.0 mm. HEAD: Distinctly narrower than the thorax, broadened behind the eyes; light to dark brownish rufous, tips of the mandibles piceous to black; finely rugose, smoother about the eyes, most rugose between the eyes and the mouth, not densely hairy, almost naked on the front. Antennæ pubescent, with 14 or 15 segments, the second almost globose, the third one-third longer than the fourth, the last almost twice the length of the preceding. THORAX: Light to dark brownish rufous; mesonotum punctate and coriaceous, not densely hairy; parapsidal grooves distinct, complete or incomplete; median groove very short or lacking; anterior parallel lines fine, distinct; lateral lines moderately broad, smooth, naked; scutellum somewhat longer than wide, well rounded posteriorly, rugose, hairy, the anterior depression rugose, smoother laterally, forming more or less distinct foveæ; pronotum laterally rugoso-punctate, hairy; mesopleuræ rugose dorsally, mostly smooth and finely rugoso-punctate, scatteringly hairy. ABDOMEN: Smooth; finely and microscopically punctate posteriorly, naked except for patches of hairs latero-basally, hairs on the edges of the posterior segments, and on the ventral spine and valves; longer than broad, somewhat produced dorsally; second segment covering not more than half the area, edges of segments oblique, well rounded ventrally; ventral spine rather long, slender. LEGS: Including the coxæ, uniform in color with the rest of the body; tarsi darker; rather densely hairy; tarsal claws strong, toothed. WINGS: Slightly tinged yellow, most so in the radial and

cubital cells; anterior margins rather short-ciliate; veins brown; areolet moderately large; cubitus not reaching the basalis; radial cell moderately short, open; the second abscissa of the radius somewhat curved, the tip parallel with the margin for a short distance; first abscissa sharply angulate at almost 90°, with hardly any projection. LENGTH: 3.0-5.0 mm.

GALL.-Large, globose to reniform, smooth, twig gall. Monothalamous, or polythalamous, containing from one to a dozen larval cells. Irregularly rounded, ovate, ellipsoid to reniform, all sizes, up to 65. mm. in diameter by 113. mm. long; the surface smooth, sparingly, shallowly roughened; reddish to yellow and brown. Internally more or less filled with compacted, soft, crystalline, brown, yellow, or white material, woody from the point of attachment to above the center of the gall; the larval cells oval, averaging 4.-5. mm. long by 3. mm. wide, more or less radiantly in the woody tissue, toward the center of the gall. Laterally, on twigs of several white oaks.

RANGE. From the Mexican border of California to Washington.

The most apparent differences between the varieties of this species are those of color and size. That we may have underestimated the constancy of these characters with some species is indicated by the absolute correlation of these differences with the physiologic differences and geographic distribution.

An astounding structural characteristic of one of the varieties is the complete parapsidal groove; the other varieties have the groove terminating very distinctly midlength of the thorax.

The occurrence of variety californicus upon two hosts without consequent variation, and the possible occurrence of fructiformis upon two more distinct hosts (as discussed under that variety), offer data concerning the non-influence of hosts upon the insect, and further confirm the idea that the gall is specific for the insect producer, without evidence of direct host influence.

The most important problem which this species may illuminate is that of the nature of alternation of generations. As discussed for each of the varieties, the northern forms quite certainly have alternate generations, but in the southern form the close concurrence of emergence dates and the appearance of fresh galls make it seem possible that there is no alternation of generations in the perpetually warm climate of southern California. If this is demonstrated it will confirm my opinion previously expressed (1920, Bull. Amer. Mus. Nat. Hist., XLII, p. 372) that alternation of generations

is an extreme development of seasonal dimorphism. With such large adults, and monstrous and abundant galls as this species has, it should not be a difficult matter for some one on the field to experimentally discover this life history.

Andricus californicus variety californicus (Bassett)

Cynips Q. Californica Bassett, 1881, Can. Ent., XIII, p. 51. Riley, 1881, Amer. Nat., XV, pp. 402, 403.

Andricus californicus Mayr, 1881, Gen. gallenbew. Cynip., p. 28. Ashmead, 1885, Trans. Amer. Ent. Soc., XII, pp. 295, 304. Ashmead in Packard, 1890, 5th Rpt. U.S. Ent. Comm., p. 110. Dalla Torre, 1893, Cat. Hymen. Cynip., II, p. 81. Dalla Torre and Kieffer, 1902, Gen. Ins. Hymen. Cynip., p. 62. Kellogg, 1914, Amer. Ins., p. 472, fig. 665. Thompson, 1915, Amer. Ins. Galls, p. 8. Felt, 1918, N.Y. Mus. Bull., 200, p. 62.

Andricus (Callirhytis) californicus Ashmead, 1885, Trans. Amer. Ent. Soc., XII, p. 294. Ashmead in Packard, 1890, 5th Rpt. U.S. Ent. Comm., p. 105.

Callirhytis californica Ashmead, 1887, Trans. Amer. Ent. Soc., XIV, p. 132.

Callirhytis californicus Beutenmuller, 1904, Bull. Amer. Mus. Nat. Hist., XX, p. 25.

Andricus quercus-californicus Dalla Torre and Kieffer, 1910, Das Tierreich, XXIV, pp. 531, 803. Trotter, 1910, Boll. Lab. Portici, V, p. 111. Fullaway, 1911, Ann. Ent. Soc. Amer., IV, p. 346; 1912, Journ. N.Y. Ent. Soc., XX, pp. 275, 280.

FEMALE.--Shows the following characters in addition to those common to all varieties of the species: Color almost entirely light brownish rufous; antennæ slightly darker than the head, darkest midlength, rather short, distinctly enlarged terminally, with only 14 segments; parapsidal grooves moderately broad at the scutellum, becoming finer anteriorly, not extending much more than half the length of the mesothorax; median groove lacking; scutellum with the anterior depression rugose, smoother laterally, but hardly forming distinct foveæ; length 4.0-5.0 mm.

GALL.-Differs from the galls of other varieties of the species as follows: Color light brownish tinged slightly reddish, to light straw yellow, usually yellow, weathering dark brown to black; usually more completely filled with the crystalline material than in other varieties. Very young galls cause a lateral swelling of the stem which often becomes 4. mm. or more high before the bark breaks; young galls are very succulent, red when first breaking thru; old galls persist on the trees for years. Maximum length observed, 113. mm. On Quercus lobata and Q. Douglasii.

RANGE.-Between the ranges of the other varieties. California: Grant (Trotter); Exeter, Three Rivers, Merced Falls, El Portal, Paso Robles, Byron, Oakdale, Gilroy (Redwood School), Palo Alto, Redwood