THE CANADIAN ENTOMOLOGISr. 357 



The Melastomataceous type includes, besides the Melastomataceae, 

 only Storckiella of Leguminosse and Mixximilianea and Amorenxia of the 

 Bixacese. 



In this type there is the same conspicuous, patent corolla and elon- 

 gate, basifixed anthers as in the Solanum-Cassia type. Dehiscence by 

 pores seems to be more specialized, since, in nearly all the forms, the 

 anthers open by a single and usually minute terminal pore instead ot two, 

 as is commonly the case in other groups. The essential difference be- 

 tween this and the Solanum-Cassia type is the elongate filaments. These 

 make possible the highly-organized anther of the Melastomatacere. To 

 this class belong clearly all of the i6r genera of the Melastomataceae, 

 except 12, in which the anthers open by more or less longitudinal slits. 



A few genera in which the anthers exhibit more or less terminal 

 poriform openings do not fall into any of the seven classes recognized. 

 Our knowledge of these forms is not sufficient to justify establishing 

 special classes for them or placing them in any of the groups already 

 recognized. They have simply been designated as aberrant forms and 

 like the first four classes, left out of consideration in the distributional 

 tabulations. 



(To be continued). 



NOTES ON NEW PHILIPPINE HYMENOPTERA. 



BY ROBERT E. BROWN, S. J., MANILA OBSERVATORY. 



Dr. W. H. Ashmead, of the National Museum, Washington, D. C, 

 has published from time to time in the Canadian Entomologist de- 

 scriptions of new genera and new species of Hymenoptera from the 

 Philippine Islands. As the great majority of these new insects were 

 either collected or bred in the garden attached to the Manila Observatory, 

 perhaps the life histories of some of them will not be without interest and 

 value. 



Microplitis Philippmensis, sp. n., Ashm. Fam. Braconidre. — Bred 

 in the Manila Observatory. This hymenopteron was bred for the first 

 time from the larva of the Sphinx moth, Chosrocampa oldanlandice, Fab., 

 but later on it was bred from three other species of Sphinx, viz.: Meto- 

 psihis adeiis^ Cram., Chcerocampa celerio, Linn., and Panacra mydon^ 

 Walker. It would appear from many observations that the parasite only 

 attacks the larvae in their earlier stages, for we have never found a 

 full-grown Sphinx larva parasitized by M. Philippinensis. When the 



October, 1905. 



