No. 2368. AMERICAN SUBTERRANEAN OALLS ON OAK— WELD. 213 



tains at the west end of Lake Chapala above San Pedro in the State 

 of Jalisco, Mexico. The fresh galls were only partly grown and al- 

 though solid, were easily cut with a knife. They were grayish in 

 color and the surface quite smooth, usually terminal on branches of 

 large spreading trees. The old galls are harder, darker, and rougher 

 and seem to persist on the tree for years. The largest gall seen 

 was given to the writer by Mr. Dwight R. Furness who collected it 

 near Ocotlan, Jalisco. It measures 16 by 12 cm. and is thus about 

 twice as large as the specimen before Doctor Ashmead, which he 

 called "the largest oak gall in the world." This species is here in- 

 cluded merely because it has been erroneously considered to be a 



root gall. 



Genus CALLIRHYTIS Foerster. 



As here treated this genus is distinguished from Andricus Hartig 

 by the absence of a tooth on the tarsal claw, thus following the in- 

 terpretation of Mayr and the European writers. Ashmead in his key* 

 reversed this interpretation, but a specimen of Andricus trilineatus 

 Hartig, the genotype of Andricus, sent by Mayr to the United States 

 National Museum, when dissected and parts mounted in balsam 

 shows a distinct tooth on the tarsal claw, proving Ashmead in error. 

 Whether the genotype of Callirhytis had a tooth on claw or not is not 

 definitely known. 



As thus understood, this large genus contains a very varied assort- 

 ment of species in which further study will probably segregate cer- 

 tain groups under new names. Even in the few species here treated 

 the first three form a group quite distinct from the others. 



KEY TO THE ROOT-GALL-FORMING SPECIES. 



1. Head and thorax uniformly covered with dense pubescence as in Disholcaspis . 2 

 Pubescence dense only on local areas on head or thorax, or sparse and not hiding 



sculpture, or bare 4 



2. Antennae 16-17 segmented, first and third subequal, California 



hartmani Weld, p. 214. 

 Antennae 13-15 segmented, third as long as first and second combined 3 



3. Pubescent area on sides of second abdominal tergite extending almost to hind 



margin. Antennae 13-segmented with last nearly twice preceding or 14- 

 segmented and last only slightly longer than preceding. corallosa Weld, p. 216. 



Pubescent area on second abdominal tergite reaching only two-thirds way to 

 hind margin. Antennae 14-segmented with last twice preceding and some- 

 times incompletely subdivided maxima Weld, p. 217. 



1. Mesoscutum shining, alutaceous or coriaceous but no part rugose, with or with- 

 out setigerous punctures 5 



Mesoscutum more or less rugose at least on some part and if coriaceous dull 9 



5. Scutellum disk almost flat, the septum between pita as broad as a parapside and 

 lying in the same plane as the mesoscutum which is unusually flat or low- 

 arclied. Head seen from above stout, its length fully half its width 6 



Scutellum disk normally convex, the septum between pits not broad or in same 

 plane as mesoscutum which is normally arched. Head seen from above more 

 lunate, its length less than half its width 7 



» 1903, Psyche, vol 10, pp. 154-5. 



