1920] St. John,— Elodea in New England 21 
Through the careful observations of several workers we now know 
that the other North American species have one type of flowers 
identical in structure with the only type produced by E. canadensis 
Michx., that is pistillate flowers bearing 3 staminodia. This was 
demonstrated for the species occidentalis by R. W. Woodward,! 
for Nuttallii by Nuttall,? and for Planchonit by Wylie.? "Thus in 
the structure of its floral parts E. canadensis stands apart from these 
other species only in its total lack of any distinct staminate flowers. 
We have only Michaux’s statement that his plant had three terminal 
cordate anthers borne on thick filaments (^TRIANDRIA * * * 
filamenta crassa: antherae terminales, cordatae"). This leaves us 
with a question that can only be settled by field work in the streams 
and ponds of the limy regions from Quebec to the mountains of 
Virginia, westward to Minnesota and southern Saskatchewan. Does 
Elodea canadensis ever produce perfect flowers, or are they always 
pistillate flowers with 3 filaments or staminodia? 
In 1814, Pursh described! a Serpicula occidentalis. He bases this 
in part on Elodea canadensis Michx., quoting Michaux’s name as a 
synonym and giving the reference to Michaux’s Flora. In the de- 
scription there is an element that tallies exactly with Elodea canadensis 
as described by Michaux from the rivers of Canada: the flowers 
being perfect, with three stamens, with strap-shaped reflexed biparted 
stigmas. But, to this Pursh adds, “foliis ternatis linearibus acutis. 
* * * In stagnant waters, frequent; from Canada to Virginia. 
A. July. v.v. * * *. Michaux describes the leaves to be oblong 
and obtuse, which is only the case in the early part of the season: 
at flowering time they certainly are long, linear, and acute.” It will 
be apparent that Pursh had himself seen a species of this genus, 
probably in Virginia, one which even at flowering time had long 
linear acute leaves. There is such a species, and it is easily dis- 
tinguished from Elodea canadensis Michx. on these vegetative char- 
acters, as well as floral characters which Pursh does not mention, 
drawing his description of the floral parts from Michaux and his 
E. canadensis. This linear-leaved species has a more southerly 
range, not being known from Canada. It occurs in southern Maine, 
1 RHuopona xxi. 219 (1919). 
? Nuttall, Thomas, Gen. N. Am. PI. ii. 242 (1818). 
3 Iowa State Univ. Nat. Hist. Bull. vi. 49 (1913). 
4 Pursh, Frederick, Fl. Am. Sept. i. 33 (1814). 
