6 



SCIENCE 



[N. S. Vol. LI. No. 1305 



the vascular structures in their minutest de- 

 tail, and loses sight of other important factors 

 in any evolutionary succession. 



Apparently no one, as yet, has taken all the 

 results from all fields of investigation, and 

 given us the result of the combination. In 

 other words, in phylogeny, we have had single 

 track minds. This has been necessary for the 

 accumulation of facts, but ujifortunate in 

 reaching conclusions. 



This is but a picture of botanical investi- 

 gations in general as formerly conducted; and 

 it seems obvious that cooperative research will 

 become increasingly common as cooperation 

 is found to be of advantage. 



The second situation in which cooperative 

 research will play an important role is less 

 important than the first, but none the less 

 real. 



It must be obvious to most of us that our 

 literature is crowded with the records of in- 

 competent investigations. Not all who de- 

 velop a technique are able to be independent 

 investigators. They belong to the card cata- 

 logue class. They are not even able to select 

 a suitable problem. We are too familiar with 

 the dreary rehearsal of facts that have been 

 told many times, the only new thing, perhaps, 

 being the material used; and even then the 

 result might have been foretold. It is un- 

 fortunate to waste technique and energy in 

 this way; and the only way to utilize them is 

 through cooperative research, for which there 

 has been a competent initiative, and in the 

 prosecution of which there has been a suitable 

 assignment of parts. In my judgment this is 

 the only way in which we can conserve the 

 technique we are developing, and make it 

 count for something. I grant that the prod- 

 uct of such research is much like the product 

 of a factory, but we may need the product. 



In one way or another, cooperative research 

 will supplement individual research. Individ- 

 uals, as a rule, will be the pioneers; but all 

 can not be pioneers. After exploration there 

 comes cultivation, and much cultivation will 

 be accomplished by cooperation. 



3. The most important feature that will be 

 developed in the botanical investigation of the 

 future is experimental control. Having rec- 



ognized that structures are not static, that 

 programs of development are not fixed, that 

 responses are innumerable, we are no longer 

 satisfied with the statement that all sorts of 

 variations in results occur. We naust know 

 just what condition produces a given result. 

 This question as to causes of variable results 

 first took the form of deduction. We tried to 

 reason the thing out. 



A conspicuous illustration of this situation 

 may be obtained from the history of ecology. 

 Concerned with the relation of plants to their 

 environment, deductions became almost as 

 numerous as investigators. Even when ex- 

 perimental work was begun, the results were 

 still vague because of environment. Finally, 

 it became evident that all the factors of en- 

 vironment must be subjected to rigid experi- 

 mental control before definite conclusions 

 could be reached. 



What is true of ecology is true also of 

 every phase of botanical research. For ex- 

 ample, I happen to be concerned with mate- 

 rials that showed an occasional monocotyle- 

 donous embryo with two cotyledons, while 

 most of the embryos were normal. The fact 

 of course was important, for it connected up 

 Monocotyledons and Dicotyledons in a very 

 suggestive way, and also opened up the whole 

 question of cotyledony. Important as the 

 fact was, much more important was the cause 

 of the fact. We could only infer that certain 

 conditions might have resulted in a dicotyle- 

 donous embryo in a monocotyledon ; but it was 

 a very unsubstantial inference. That problem 

 will never be solved until we learn to control 

 the conditions and produce dicotyledonous em- 

 bryos from Monocotyledons at will, or the re- 

 verse. Comparison and inference must be re- 

 placed by experimental control; just as in the 

 history of organic evolution, the method 

 shifted from comparison and inference to ex- 

 perimental control. It will be a slow evolu- 

 tion, and most of our conclusions will con- 

 tinue to be inferences, but these inferences 

 will eventually be the basis of experiment. 

 In fact, most of our conclusions are as yet 

 marking time until a new technique enables 

 us to move forward. 



These illustrations from ecology and morph- 



