592 



and nearly or quite entire. In what I have supposed to be laciniata 

 the leaves are rather ovate or deltoid, and all nearly similar. The fourth 

 seems to have been hitherto unnoticed ; and if it should prove distinct 

 from the second, I would call it campanulata. But the form of the 

 sepals varies so much in these plants that I dare not place much re- 

 liance on it. 



The seeds in all the plants of this set are similar, brown, com- 

 pressed, and of a waxy surface. 



I will conclude with a few observations on the sepals of the fertile 

 flowers, for I cannot help thinking that more extended and more 



Sepals of Atriplex as specified below. 



exact observations will at length show some definite limits to the va- 

 riations in their forms, and enable us to make more use of them in the 

 specific characters. There is a continued, though not indeed a very 

 regular progress from ovate to hastate or triangular, to rhombic, to a 

 square placed diagonally, and thence to campanulate, a slightly waving 

 line on the two sides of the square or rhomb forming an easy transition 

 to campanulate with an extended termination. In the figures on the 



