1917] Fernald,— New Saxifragas 143 
slightly elinpnte; but the flowers, as in S. nivalis, are sessile or subsessile 
in glomerules and not slenderly pedicelled as in var. labradorica. 
SAXIFRAGA VIRGINIENSIS Michx. In Essex County, Massachusetts, 
this species seems peculiarly subject to minor variations, four fairly 
distinguishable forms of the species there being found. In typical 
S. virginiensis the petals are white and the flowers and fruits pedicelled, 
the cymose branches of the panicle elongating in fruit. In the town 
of Andover, however, a singular form of the plant occurs; in fact, the 
only collections (3 in number) seen from Andover all belong to this 
peculiar form which may be called 
S. VIRGINIENSIS, forma glomerulata, n. f., floribus sessilibus in 
glomerulis dispositi 
Flowers ge. i in glomerules at tips of the branches.— Massacuu- 
sETTs: Andover, 1901, A. S. Pease, no. 672; ledges, Rabbit Rock 
Pond, Andover, ‘April 27, 1902, A. S. Pease, no. 673; Prospect Hill, 
Andover, May 24, 1902, A. S. Pease, no. 671 (TYPE i in Herb. New 
En gland Botanical Club). 
In its sessile flowers borne in glomerules at the tips of the branches 
- this singular form from Andover strongly suggests S. nivalis and by 
ordinary treatments of the genus in which S. nivalis is separated 
chiefly by its sessile flowers the Andover plant would fall readily into 
that species. S. virginiensis is distinguished, however, from S. nivalis 
by a number of characters which have not always been clearly recog- 
nized. In S. nivalis the bracts are from one-half to fully as long as 
the branches of the inflorescence; the calyx-lobes spreading in fruit; 
the petals 1.5-3.5 mm. long, about equaling or only slightly exceeding 
the calyx-lobes. In S. virginiensis, on the other hand, the bracts are 
many times shorter than the branches of the inflorescences; the 
calyx-lobes ascend in fruit; and the petals in the normal forms of the 
species are 4-5.5 mm. long, 2-3 times as long as the calyx-lobes. 
Some aberrant forms of S. virginiensis, however, occur with the 
petals very small or wanting. These plants, of which two have been 
described as varieties, are not, however, varietal in character, seeming 
to occur merely as aberrant colonies within the range of the typical 
S. virginiensis and they should be treated rather as forms. The first 
is 
S. virGINIENSIS, forma chlorantha (Oakes), n. comb. 8. vir- 
giniensis, var. chlorant akes in Hovey’s Mag. xiii. 218 (1847), 
strato as follows: “Petals pale green, instead of snow white, as in 
