692 MR JOHN MCLEAN THOMPSON ON 



It is by this method that a floral receptacle on which arise floral parts, in an order 

 departing slightly from an acropetal succession, has been shown to have been recently 

 represented in the ancestry by a receptacle to which a strictly acropetal order of 

 floral-part development was peculiar. 



Support for such an opinion is sought in the existence of flowers with acropetally 

 arising floral-part-primordia, among forms held to be of near affinity to those dealt with. 



There are here two assumptions : first, that the organisms, with which present 

 forms are compared, existed ; and second, that even if they did exist, the organisms now 

 dealt with were derived from them. 



These assumptions may often have been unavoidable, but the opinions formed have 

 as a rule been accepted only when suj)ported by the results of comparisons between 

 living organisms — and, where possible, between fossils — believed to be closely related 

 to the organisms immediately considered. 



Still, the element of uncertainty remains, and cannot be greatly reduced until a wide 

 knowledge is obtained of the actual development of each systematic character. 



The examination of a single character possessed by a group of organisms already 

 systematised shows that, while in some members of the group the character is well 

 marked, in others it is but little developed. Should any of the organisms possess the 

 character in so slight a degree as to off"er an opportunity for the study of the actual 

 initiation of the character, the observer is provided with the best material available 

 with which to begin a developmental study. The outcome of such a study will be 

 a knowledge of details, the determination of the point at which the character appears, 

 and, if possible, a recognition of the consequences of its initiation in the organism. 

 It does not follow that the initial step in the evolution of the character in the ancestry 

 of the related organisms was identical with that described, but the observer is at least 

 provided with actual facts with which others can be compared. 



It has seemed advisable to prefix these remarks to the following observations, which 

 are intended to form an introduction to detailed statements of the evolution of floral 

 zygomorphy in various affinities. 



It has been customary to consider the Angiospermic flower from the point of view 

 of Symmetry, for it has seemed that a knowledge of the degree of symmetry of flowers 

 is necessary to the proper conception of floral mechanism. 



An examination of any flower from this standpoint may be guided by a recognition 

 of the importance of definite stages in development of the mechanism, and should be 

 directed by a clear conception of the function, and time of functioning, of each part. 



A calyx may be protective during the bud stages, and may be shed when the flower 

 has expanded. In some cases sepals perform a double function. In the bud they are 

 protective, but they persist, and become petal oid in the expanded flower. The outer 

 perianth-segments may have fallen, and the petals may have been expanded for some 

 time before pollen is liberated from the anthers, or the stigma is receptive. 



At what stages, then, is the symmetry of a flower to be determined ? The answer 



