Proceedings of the Club 213 



in their rachis. Still another is the enlargement of one perigone- 



segment, to act as a scale, and in its absence, to protect the flower 

 of a catkin, spike or other dense inflorescence, as seen in some 

 species of Eriocaidon. ' ' ■ 



Other forms of irregularity, those with which the paper had 

 properly to deal, were classified and shown to represent different 

 degrees of modification. The lowest form was regarded as the 

 simple curving upward of a horizontal or declined androecium or 

 gynaecium, as seen in Mirabilis. The next involved an accom- 



- 



panying curvature of perigone, as in Cyrtanthus, then successively 

 an oblique base or mouth, as in ma|y Gesneriaceae, the distinction of 

 the tube in varying degrees, as seen in the same family, the exag- 

 geration or reduction of the anterior or posterior portion of the 

 limb, as in Chioscographia or Pteropetalon, and in bilabiate corollas, 

 the arrangement of such corollas to form radiant inflorescences, as 

 in many Umbelliferae, variations in size as well as direction of an- 

 terior and posterior stamens, as in Cassia, and numerous forms of 

 appendaging. , 



These forms were traced among the Monocotyledons and Dico- 

 tyledons respectively. None were found among the 2 1 lowest of the 

 43 families of Monocotyledons. Of the io next highest, 5 show 

 none, 4 show slight or doubtful forms, while the highest, Liliaceae, 

 with 1 97 genera, twice as many as the other 9 families combined, 

 shows, amidst general regularity a few highly irregular genera, 

 two of them simulating Orchidaceous forms. Of the 12 highest 



families, only 3 are regular. Five of the highest 6 are very irreg- 

 ular indeed, the highest, Orchidaceae, phenomenally so. It thus 

 appears that an increased tendency to irregularity is indicative of 

 higher development, but it is liable to occur in families and groups 

 of families usually distinguished for its absence. 



This principle was then shown to be even more clearly illus- 



- 



trated by the Dicotyledons. In the 53 lowest families, but 4 show 



■ 



irregularity. Only 1 of these is found among the first 39, and 

 this is Aristolochiaceae, with a single irregular genus. Among 

 the next 120 families, 27 show irregularity, and these are rather 

 uniformly distributed among the others. Then come 19, several 

 showing slight irregularity and one very irregular indeed. The 

 next 17 are, with one exception, highly irregular, one of them 



