74 



In the female heads the scales are larger than in the male 

 heads, and form a well marked calyx, from four to seven being ^ 



attached to a single flower. The normal number of pistillate 

 scales is four or five, corresponding with the number of the angles 

 on the fruit, but often two or three others occur outside of these. 



The pollen consists of single grains which are globular, with 

 a smooth surface, and very similar to those of Typha angustifolta, 

 being in 5. eiirycarpiim iwd to ^w inch in diameter, and somewhat 



smaller in S. simplex. 



The stigmas are unilateral, lying along the upper portion of 

 the style for a considerable distance from the apex. When the 

 style is cleft, as Is often the case, both divisions bear stigmas on 

 the ventral side, and when the style is simple the stigmatic por- 

 tion is on the upper side. The simple styles as they project from 

 the heads curve slightly upwards so that the pollen falling from 

 the anthers is more easily caught by the stigmas. 



Most of the species bear double stigmas, either frequently or 

 occasionally, and, of course, in such the ovaries are two-celled and 

 two seeds are formed. One of them, however, is generally sup- 

 pressed by abortion so that the perfected fruit is single. The 

 number of stigmas, therefore, is not a safe character to use in 

 specific distinctions, except when taken in connection with other 

 characters. The length of the stigmas, also, is variable, even 

 in the same form, and but little reliance can be placed upon it in 

 classification. The fruit-characters must form our chief ground 

 of distinction in any accurate determination of the species. 



These points should be kept In mind in estimating the attempt 

 made in this paper to classify the members of this genus compre- 

 hensively. I am not aware that any effort has been made to do 

 this heretofore. Dr. Geo. Engelmann, whose lamented death has 

 left a large void in our list of American naturalists, divided most 

 of the North American species into groups for the 5th edition of 

 Gray's Manual; but his division is, of course, limited, and 

 founded upon subordinate characters. The summary now pre- 

 sented is based upon an examination of our North American 

 forms, all of which, so far as known, have been carefully studied. 

 Some of the foreign species, however, are not contained or not 

 well represented in any trans- Atlantic herbarium, and conse- 



