832 



SCIENCE 



[N. S. Vol. XXXI. No. 804 



The writer ventures to offer the following 

 conjectures of relationship with the hope that 

 cultures may some day prove their correct- 

 ness. They may turn out eventually to show 

 only the errors which the method of analog- 

 ical inference may lead to, but they serve 

 well to illustrate its application to this sub- 

 ject and are offered with the hope that they 

 may have some value. 



Roestelia hyalina on a species of hawthorn, 

 Cratwgus, is a peculiar form which has been very 

 little known up to the present. So far as the 

 writer can make out there is no published record 

 of any but the original collection from South 

 Carolina made in 1860. The writer has recently 

 rediscovered it on some herbarium specimens of 

 the host plant at the New York Botanical Garden 

 and the Arnold Arboretum so that its occurrence 

 is after all not so rare; its distribution is now 

 known from North Carolina to Florida. This 

 adds new zest to the attempt to trace out its 

 alternate phase. Morphologically R. hyalina has 

 two very striking characters; first, the entirely 

 smooth walls of the peridial cells, and second, the 

 small wart-like protuberances on the leaves in 

 which the peridia are borne. Only one other 

 known species, B. Botryapites on Amelanchier, 

 the service-berry, has these characters and it has 

 seemed to the writer for some time that these 

 two forms must be related to similar telial stages. 

 R. Botryapites is known to be connected to Gym- 

 nosporangium biseptatum on the white cedar, 

 Chamwcyparis thyoides. O. Ellisii is another 

 white cedar rust, similar to O. biseptatum in the 

 form of the distortion produced on the host and 

 in the character of the spores, both having 2—4- 

 celled teliospores. G. Ellisii has been supposed 

 by some to be connected to Rcestelia transform-ans 

 on Aroma arbutifolia, but a careful examination 

 of the culture record shows that this conjecture 

 has never been successfully demonstrated. On the 

 other hand, there are so many negative results 

 that it seems almost safe to conclude that it has 

 been disproved. It seems very probable that since 

 one of these two forms of Rcestelia, forms which 

 are in a class by themselves on account of their 

 smooth peridial cells and external anatomy, be- 

 longs to a white cedar rust of a certain type, that 

 the other may belong to the only other white 

 cedar rust of the same general type at present 

 known. Recent collections of G. Ellisii by Stone 

 in Alabama and Tracy in Florida make its known 

 range from Massachusetts to Alabama so that it 



is quite feasible to suppose it connected with a 

 form which is known in the heart of that range, 

 North Carolina to Florida. 



If an hypothesis provisionally formed 

 either by association or analogy can be sup- 

 plemented by inferences drawn from homology 

 it is very materially strengthened. Homol- 

 ogy might be defined so far as its application 

 to botany is concerned as the morphological 

 likeness existing between elements which may 

 have become adapted to quite different func- 

 tions. In applying this to the subject under 

 discussion, for instance, if there is an essen- 

 tial structural resemblance between the aecio- 

 spores and the urediniospores of a species 

 they may be said to be homologous. Some 

 notable examples in which homology in this 

 sense has assisted in detecting genetic rela- 

 tionships have already been recorded by Dr. 

 J. C. Arthur in his first report of " Cultures 

 of Uredinese '" and may be mentioned here. 



Field observations had suggested that Puc- 

 ciiiia Yilfw on a grass, Sporoholus longifolius, 

 was related to ^cidium verhenicola on Ver- 

 hena. It was found that the closer the Ver- 

 hena plants stood to tufts of the rusted grass 

 the more thickly they were covered with aecia, 

 and that the plants some distance away were 

 entirely free. This is a good example of the 

 working of the law of association. Before 

 cultures were made, however, a resemblance 

 in form was observed between the seciospores 

 of ^cidium verhenicola and the uredinio- 

 spores of Puccinia Vilfce. The two sorts of 

 spores were similar in shape and surface 

 markings, and both had colorless walls much 

 thickened at the apes. Later successful cul- 

 tures proved that this homology was not a 

 mere accident in this case and suggested that 

 it might be the sign of relationship in other 

 instances. During the same year a similar 

 morphological correspondence was found be- 

 tween the £eciospores of an ^cidium on 

 Fraxinus, the ash tree, and the urediniospores 

 of a Puccinia on Spartina, cord-grass, and 

 with this as the only clue cultures were at- 

 tempted. They were successful and thus 



=1 " Cultures of Uredinese in 1899," Bot. Gaz., 

 29: 274-275, 1900. 



