652 



SCIENCE 



[N. a Vol. XXXIX. No. 1009 



Gymnoconia, Phragmidium, Phragmopyxis, 

 Blastospora, Rostrupia, Triphragmium, Hapa- 

 lophragmium, Sphwrophragmium, Anthomyces, 

 Uromycladium. These are distributed among 

 the three subfamilies, Phragmidiese, Uro- 

 pyxidese and Puccinieas, into which the family- 

 is divided. The limitation placed on these 

 subfamilies has not been very rigid, for the 

 genus Triphragmium is included in two of 

 them, the Phragmidiese and the Puccinies. 

 The authors, however, state that they are un- 

 certain regarding the place which Triphrag- 

 mium, Hapalophragmium and Sphwrophrag- 

 mium should occupy in their key. The remain- 

 ing genera of the family Pueciniaceas are 

 Dicheirinia, Qerwasia, Hemilea, Ravenelia, 

 Neoravenelia, Euehneola, Pucciniostele, 

 STcierha, and presumably we may expect the 

 next part to deal with them in the order given. 

 The classification shows conservatism on 

 every hand and especially in the selection of 

 the characters upon which it is founded. The 

 old idea of the importance of the teleutospore 

 is maintained. Such a method can be made 

 to work very well as long as only the common 

 things are considered from a " practical stand- 

 point," but when all forms are considered from 

 a scientific standpoint it can not be said to 

 have much in its favor. The result in the 

 present key is uncertainty and lack of uni- 

 formity. The attempt to arrange the key in 

 such a way as to show relationships of the 

 genera is highly desirable, but is not attained 

 when the boundaries for genera are so loose 

 that species contained within them admittedly 

 indicate relationships to different subfamilies. 

 But the composite character of genera can not 

 be avoided with the one character scheme as 

 a basis. for grouping. Neither can the segre- 

 gation of closely related forms be prevented 

 as long as this system is maintained. The 

 purely artificial character of number of cells 

 in the teleutospores will throw forms which 

 are of undoubted relation into different genera. 

 Numerous examples illustrative of this fea- 

 ture have already been cited in the literature'^ 

 and more are constantly being found as care- 



1 Arthur, Mycologia, 4: 54-56, 1912, and Orton, 

 Mycologia, 4: 194-204, 1912. 



ful comparisons are made. An arrangement 

 which places the peach and plum rusts in the 

 genus with Puccinia graminis and then sepa- 

 rates a few forms from Puccinia into the 

 genus Rostrupia may be " practical," but if it 

 is any way natural it must be accidental. If 

 Rostrupia which differs from Puccinia only in. 

 having the teleutospores with 3 or 4 cells 

 should be maintained, it is not clear why sev- 

 eral of the species of Gymnosporangium which 

 have more than 2 cells should not be separated 

 into a genus by themselves, or why the old 

 genus Phragmidium should not be broken, up 

 into several genera, since the number of cells 

 in the teleutospores in this group is highly 

 variable. If the forms on Rosaceous hosts 

 which have 2-several cells are worthy of 

 generic standing outside of the ordinary 

 Puccinia and Rostrupia genera, then it is not 

 clear why the 1-celled forms on these hosts 

 should not be separated from Uromyces, but 

 such has not been done. 



The bulk of the part is taken up by the 

 genera Gymnosporangium and Phragmidium,, 

 136 out of 192 pages being devoted to them 

 and divided nearly equally between them. The 

 monograph of Gymnosporangium is of inter- 

 est in comparison with the one published by 

 the writer about a year earlier as a Bulletin 

 of the New York Botanical Garden.^ The 

 order of arrangement of the species, the plan 

 of the keys, and the form of the descriptive 

 accounts are the same as introduced in the 

 writer's bulletin. Three additional species 

 are included in the Sydow monograph, all 

 described since the appearance of the writer's 

 publication and founded on material not seen 

 by him. As to the validity and relationship of 

 the species there has not been the slightest 

 disagreement. With one or two exceptions it 

 is also to be noted that the specific treatment 

 of the Phragmidium group is identical, so far 

 as North American species are concerned, with 

 Arthur's account in the TJredinales, " North 

 American Flora." The latter, however, refers 

 some of the species to other genera, Earlea, 

 Euehneola, while the Sydows refer all to the 

 genus Phragmidium. 



As regards certain nomenclatorial questions, 



2 Vol. 7, No. 26, 1911, pp. 391^83. 



