Arthur: Nineteen Years of Culture Work 13 



which that form undergoes in passing through its whole life 

 cycle. Many rusts are commonly collected in only one or two 

 stages of their development, or the several stages are taken as 

 independent objects, and to grow such rusts so as to keep them 

 under direct observation and be able to note the succession of 

 stages seemed highly desirable, and especially so for the heteroeci- 

 ous species which pass their gametophytic and sporophytic stages 

 upon wholly unlike and unrelated hosts. It was natural, there- 

 fore, to direct chief attention, especially at first, toward unravel- 

 ing the tangle of heteroecious forms. 



Nobody knew how many rusts were to be found on the North 

 American continent and its islands. There were possibly a thou- 

 sand or more names in existence, but how many of these names 

 applied to single and independent life cycles, and how many to 

 parts of cycles, or were synonyms, no one had attempted to say. 

 It was, in fact, only with the existing names that I had to do. 

 It was no part of my problem to discover new species, or to give 

 new names, either in preparing manuscript for the North Ameri- 

 can Flora, or in conducting cultures, except in so far as these 

 were required for the systematic development of the work. 

 Many longer or shorter excursions were made during the progress 

 of the cultures, some of them a thousand miles or more, but they 

 were all for the purpose of making field observations upon known 

 species, and in no case for making discovery of new species. 

 The new species that were found were an incidental result. 



The first year of the culture work, that of 1899, was very 

 encouraging, and developed no particular difficulties calling for 

 solution. So far custom was followed in the application of 

 names, and it had not been necessary to apply any formula to 

 decide what constituted a species. The assumption that forms 

 on the same or closely related hosts, having no striking morpho- 

 logical differences, were of one species seemed a sufficient hy- 

 pothesis, and the corollary necessarily followed that cultures would 

 show the range of hosts for each species, as well as serve to 

 demonstrate the stages and spore-forms in the life cycle. Cer- 

 tain features in connection with the common Euphorbia rust did 

 indicate that difficulty might be found in the application of the 



