May 27, 1910] 



SCIENCE 



831 



other words, might be considered a general 

 law among heteroecious forms. 



In- addition to this deductive method, 

 which requires field observations upon which 

 to base its inferences, there is another method 

 by which predictions of relationship may 

 sometimes be inductively made. In some 

 instances there is something peculiar about 

 either host or fungus, or both, which will 

 permit the formation of an hypothesis. These 

 peculiarities may be in form, habit, range 

 or other characteristics, and are usually of 

 such a nature that they may be studied out 

 in the herbarium or laboratory. This second 

 method deals very largely with the principles 

 of analogy and homology. 



It is the analogical method of inferring 

 that what is true of one species is probably 

 true of others similar to it which makes us 

 conclude, for instance, that the species of 

 Coleosporium, common rusts of the compo- 

 sites especially, are related to leaf-inhabiting 

 forms of Peridermium on pine trees, that the 

 species of CronarUum are connected to bark- 

 inhabiting forms of Peridermium on pine 

 trees, and that the Gymnosporangia, the 

 cedar-rusts, have Rcestelice on members of the 

 apple family as their secial forms. This gen- 

 eral theory for the assignment of certain 

 form-species to their telial genus has already 

 been illustrated in a paper of which the writer 

 was junior author,^ and subsequent culture 

 work has demonstrated not only the accuracy 

 of the predictions, but also the importance of 

 such theorizing. 



The writer wishes now to call attention to 

 still better examples of the application of this 

 analogical process. Perhaps the procedure 

 may be made clearer by a fuller statement of 

 the formula and the consideration of some 

 concrete examples. The examples will be 

 drawn from the group of cedar-apple fungi, 

 Gymnosporangia, because of the writer's 

 familiarity with this group. 



Analogy has been explained in this way: 

 Two things which are similar in one or more 



' See " North American Species of Periderm- 

 ium," Bull. Torrey Botan. Club, 33: 403-438, 

 1906. 



respects are of the same general type or char- 

 acter; therefore a certain proposition which is 

 true of one is likely to be true of the other. 

 In applying this to the fungi, as well as in 

 other cases, it is especially important that the 

 characters selected for comparison should be 

 fundamental ones and not merely of a super- 

 ficial nature. Some accumulated knowledge 

 in a field, even if it should only be in the 

 form of negative answers to previous conjec- 

 tures, may not be without value in forming 

 new hypotheses. 



The following example may be cited in 

 which cultures have already shown the cor- 

 rectness of an hypothesis formed by the 

 method just explained. 



Some time ago what appeared to be a true 

 Rcesielia was found upon an herbaceous plant 

 of the rose family. This was considered re- 

 markable because it had always been sup- 

 posed that all Rcestelice inhabited only woody 

 plants of the apple family. Upon thorough 

 examination, however, this was found to have 

 all of the morphological characters of the 

 roestelial forms and it was, therefore, con- 

 cluded that it was most likely associated with 

 a cedar-rust, as other members of this form- 

 genus are. There was in the range no unat- 

 tached species of Gymnosporangium known 

 which might have such a connection; this 

 discovery called, therefore, for the detection 

 of a new form. From the great resemblance 

 of this rosaceous Rcestelia to the secial form 

 of Gymnosporangium, Nidus-avis, a rather 

 common and well-known cedar rust, it was 

 predicted that the telial stage, when found, 

 would resemble 0. Nidus-avis. This new 

 telial stage has been collected and cultures 

 have been made proving the correctness of 

 the assumption as to relationship. The pre- 

 diction as to structure was also strikingly 

 fulfilled, showing that it is not only possible 

 to show the probable existence of new forms 

 by this method but even to anticipate their 

 characters. This species has been named 

 Gymnosporangium exterum and a fuller ac- 

 count of its discovery together with original 

 description and culture record may be found 

 in Mycologia, 1 : 226, 22Y, 253 and 254, 1909. 



