176 CARROLL— ON DEVELOPMENT OF 



lacaceae; in the more evolved Rosaceae and Oxalidaceae; and finally 

 in the Labiatae and Scrophulariaceae. Families lying between the 

 extremes show abundant occurrence of cleistogamy. It would seem 

 that if mutation is the sole cause of cleistogamy, it must have arisen 

 many times in these widely separated families. This multiple origin, 

 however, would not necessarily be an objection to the theory, as there 

 is abundant evidence of other plant structures appearing in divergent 

 groups of plants. 



The evidence at hand is on the whole in support of Goebel's explana- 

 tion of cleistogamous flowers as a product of reduced nourishment. 

 Darwin thought the existence in cleistogamous flowers of altered struc- 

 tures especially adapted for cleistogamous fertilization argued against 

 the idea of arrested development (12). He mentions in particular the 

 altered style in Viola and the lack of style and the open stylar canals 

 in many forms. Bennett believes the difference between the two types 

 of flowers is original, but the grounds for his opinion are not safe. He 

 described a difference between the sepals of cleistogamous and chas- 

 mogamous flowers in the earliest recognizable bud. This, however, is a 

 comparatively late stage. The arrest may be effected in the primordia 

 of the flower, or perhaps the "arrest" or alteration in vital processes 

 may take place at even an earlier stage, before the primordia have 

 appeared at all. When alteration of structure does occur in the cleis- 

 togamous flowers it is generally a matter of the reduction of parts. The 

 petals fail to develop or develop only to a reduced size, the stamens lack 

 parts as in Impatiens, or their number is reduced even to one stamen, 

 the style is lacking and stylar canals remain open, the number of ovules 

 is reduced. These are plainly differences due to loss or reduction of 

 parts. Where deviation occurs in the line of apparently added struc- 

 tures, these structures are evidently due to such growth as occurs from 

 altered primordia. The conditions which cause these deviations have 

 been determined; the morphology of the mature parts has been described 

 in a great variety of forms ; some work has been done on the histological 

 developments of parts. Further exact work is needed to determine 

 the point at which histological differentiation first occurs, and to deter- 

 mine the exact physiological processes which result in histological 

 deviation in the two types of flowers. When there is more information 

 along these lines, further attempts to explain the biological significance 

 of cleistogamy will be in order. 



