his own theories were innumerable; some of them he answered by the great investigations already mentioned, others he left unanswered for various reasons. Only about two months before his death he pointed out new lines of investigation. Among other things he said that there were many inconspicuous flowers not known to be visited by insects during the day, and the natural inference is that they are selffertilized. And he pointed out the desirability of finding out whether these flowers are visited at night by the innumerable individuals of the many species of minute moths. If they are not so visited, why do they expand at all? Why are they not cleistogamic? He suggested, as a mode of procedure, smearing the flowers with viscid matter and then looking for insect scales; but gave the caution that it would be necessary to prove that the matter employed was not in itself attractive to insects. 1 It is a fascinating study to follow out the suggestions that came to him and that he made to others, to note the various degrees of success with which the investigations were made by others, to compare the spirit and methods with which they were made with Darwin's own spirit 1 Müller, Fertilization of Flowers, Prefatory Note, by Charles Darwin. and method. He knew the importance of studying the speech of monkeys in relation to his belief in the descent of man, and expected valuable results from it. "I wish," he wrote to Asa Gray, "some one would keep a lot of the noisiest monkeys, half free, and study their means of communication.” 1 Mr. Garner has recently written magazine articles and a book on the subject, and has even visited the ape country in Africa with elaborate arrangements for studying the speech of monkeys in their native haunts. It is too early to forecast results, but this case illustrates well how different phases of Darwin's theories have attracted different types of men, and how caution or the want of it may make or break confidence in the results of their investigations. 1 Life and Letters, Vol. II. p. 183. CON XIII. ERRONEOUS DEDUCTION. The ONSIDERED simply as a logical process, deduction is no more interesting in the hands of the modern investigator than it was in the hands of the medieval schoolman. The scientist uses it, apart from its importance in proof, or effort to convince others, merely as an instrument with which to test what is known, and to develop its unknown consequences. infallibility of the process is altogether hypothetical. The conclusion is true only if the premises are true; and since the truth of the premises is oftener the matter in question than even the investigator dreams, the effort to get at new truth by anticipating the consequences of theory often results in false conclusions. This must especially be the case when there is no apparent reason to question the truth of the premises; when they would seem to have been permanently established by repeated crucial tests. In a number of instances Darwin went ERRONEOUS DEDUCTION. TEESF UNIVERSITY CALIFORNIA clearly wrong in his deductions. Some of them he corrected himself, for some he accepted the corrections of others; and some have never yet been corrected directly, but only by the adoption of the contrary principle from a study of facts similar to those on which Darwin went wrong. When his attention had become fixed upon the cowslip (Primula veris), he found that there were two forms of flowers on plants of this species. He said, "The first idea which naturally occurred to me was, that this species was tending towards a diœcious condition; that the long-styled plants, with their longer pistils, rougher stigmas, and smaller pollen grains, were more feminine in nature, and would produce more seed; that the short-styled plants, with their shorter pistils, longer stamens, and larger pollen grains, were more masculine in nature.” 1 Nothing would seem more natural than that the structural differences between the flowers should indicate differences in sexual function. The knowledge of dioecious plants and belief in the modification of species could plainly suggest but one interpretation of the structural differences of the two kinds of flowers. Either they indicated what was inferred from them, or they could 1 Different Forms of Flowers on Plants of the same Species, pp. 18-21. indicate nothing at all. Had his conclusion been left at this point, it would probably have been accepted as both interesting and quite certain. The degree of certitude with which such an inference is received depends on the number of facts involved, their relation to each other, and the degree to which they act as convergent evidence toward the one conclusion. In these respects the facts were all that could be desired. It would seem that, if it is possible to make any inferences at all concerning function from the structure of plants, it would have been so in this case. For Darwin, as for every true student of nature, deductions exist only to be verified. The indirect evidence from structure he supplemented by experiments on the actual production of seed by the two forms. He might have pointed with pride to the cowslip as a plant in an actual state of transition, as a fine illustration of his theory. But after describing in detail the differences of structure in the two forms of flower in the cowslip, he said, “The question seems well worthy of careful investigation." He made preliminary experiments which of themselves would have been conclusive, but used them to lay a basis for his much more extensive experiments; they suggested |