WHILE COLLECTING HERBARIUM. SPECIM ENS

from golden shower trees (Cassia fistula 1., a legume from India) at my home in Nuuanu Valley, Honolulu, I noticed that three of our four trees (A, C, D) bore the common form of flowers and that a fourth (B) bore aberrant forms. This discovery led to the examination of other trees of the same species. After study-ing from June to September, 1948, four or more flowers of each of 26 trees ofC, fistula in seven different localities in Honolulu, I found that trees A, C, andD definitely bore the com-mon form of flowers, which were nearly uni-form. Only tree B bore flowers deviating dis-tinctly from those borne by the other trees. Examination of all these flowers and also of literature dealing with C.fistulaled to two con-clusions: (1) The stamens of the common form of C. fistula needed redescription; and (2) the petals, sepals, and stamens of the aber-rant form of C. ftstula differed in several ways from those of the common form. COMMON FLOWERS The common form of the flower of C. ftstula has 5 petals, 5.sepals, and 10 stamens, and it has been described many times. Wight (1840, pI. 269) has a figure which more or less epitomizes these descriptions and from which is here reproduced a detail (Fig. 1), showing flowers for comparison with my drawing (Fig. 2). In Wight'S figure, the three shortest stamens appear to be alike. The petals and sepals of published descriptions seemingly were like those of flowers on 25 of the trees examined by me. The stamens, however, wer