90 PROCEEDINGS: BOSTON SOCIETY NATT'RAL HISTORY. 
purple though sometimes becoming yellowish; those of var. .salmlo- 
7ic7)ftis at first purple, but in maturity, becoming yellow like those of 
the more southern species. On Newfoundland and Cape Breton 
B. iodandra is in flower through August, l)ut the Sable Island plant 
is some weeks later, the material collected August 30, being only in 
bud. 
Combining the color, foliage, and most of the flowe "-character'^ 
of B. ''odandra with the habit and more deeply cleft calyx of B. 
paniculatn, the Sable Island plant presents an interesting transiti(m. 
B. paniculata is characteristic of the Coastal Plain from Louisiana 
to southeastern Massachusetts, rarely extending to southernmost 
Maine. B. iodandra in typical development is confined to Cape 
Breton and Newfoundland; and the transitional plant to the isolated 
Sable Island, one of the last remnants of the ancient continental 
shelf which extended from southern New England to the Newfound- 
land banks. It would tlius seem probable that the widely distrib- 
uted southern B. paniculata originally spread northward on the 
continental shelf, becoming modified toward the North, the Sable 
Island plant still retaining some distinctive ^^amcw/a/a-characters, 
which have disappeared from the more northern and further isolated 
B. iodandra of Cape Breton and Newfoundland. 
Menyanthes trif oliata L. " Quite rare, in ponds at No. 3 sta- 
tion." Found only by J. Macovn (C. no. 22,541). 
Centaurium umbellatum Gilib. Very common in the wet 
dune hollows, and by the w^et sandy borders of the fresh-water ponds. 
Not near the Life Saving Stations or the Lighthouses and not giving 
any indication of being introduced. This species has been known 
in North xAmerica for a long time, but it has universally been treated 
in botanical manuals as an introduced plant. This seems to be 
the true explanation in the greater number of the cases, such as the 
record from Concord, Massachusetts,^ A.W. Hosmer reporting it "found 
at Concord in 1890, not seen since." The species is occasional 
in the State of Michigan, but there it also seems to be an introduction. 
There is, however, a station near Oswego, New York, which has 
been known for nearly a hundred years. In 1833 Beck reported,^ 
^As Eryllraea Centaurium Pers., Rhodora i. 224 (1899). 
^Ap Erylhraea Centaurium Pers., Beck, L. C: Bot. of N. and Middle States, 
242 (1833). 
