166 PROCEEDINGS OF THE ACADEMY OF [1892. 



state of expansion when the curvature of the stipe began, we might 

 conjecture that it was some action of the pileus or cap that caused 

 the stipe's curvature, but it was evident that the departure from the 

 straight line, was nearly or quite simultaneous with the rupture of 

 the hymen, and that nothing connected with any external circum- 

 stances influencing the expansion of the pileus had anything to do 

 with the matter. 



When we consider the facts as relating to these cryptogams, and 

 the facts in connection with the various angles in the branching of 

 the same species among flowering trees it seems to be a fair inference 

 that the law which determines the direction of growth has very little, 

 if any, relation to conditions of environment. If the mystery is 

 ever solved, it will probably be found among the properties of the 

 single cell, from which the whole characteristic of the plant is 

 finally developed. 



Tricarpellary Umbellifers, 



Dr. Lindley observes (" Vegetable Kingdom ", p. 774) " in some 

 accidental cases three carpels have been found " in Umbellifers. In 

 these days when the genesis of families is an active study, departures 

 from usual characters are of more value than when Lindley wrote. 

 It is these variations that frequently give the clue to family relation- 

 ships. It is worth recording that in Eryngium planum Linn., a 

 species from the north of Europe and Asia, trigynous flowers occur 

 with some frequency. I rarely examine a head without finding one 

 such flower, and I have found six in one head. There is usually 

 one near the largest and longest involucral bract, and they are 

 generally found in the lower part of the head, in the vicinity of the 

 bracts. 



A plurality in the number of carpels is one of the characters 

 relied on to distinguish Araltacete from Umbellifers. The trigynous 

 flowers of this, and probably other Eryngiums show a line of relation- 

 ship between the two families. 



The facts may also suggest a not distant relationship with 

 Valerianaceoi. 



A MODE OF VARIATION IN StELLARIA MEDIA. 



Among a number of well developed plants of the common 

 duckweed growing on a compost heap, it was interesting to note that 

 no two seemed to be exactly alike. They differed from one another 



