Many years after this work was published, I discovered among my papers a memorandum which unfortunately I had overlooked, containing further evidence in support of the foregoing conclusion. With the omission of an error concerning the species of plant, I reproduce this memorandum just as it stood:—
“I found at Dieppe, July 1, 1860, in a garden near the sea a sample of cultivated wild flower (I thought it was grown as an ornamental flower) in which some of the single flowers of the umbel were developed into groups of flowers thus:—
“In the case where the transformation was fully effected the umbellule had six flowers, answering to the six petals of the original flowers. In other cases the transformation was incomplete. There were instances where but two of the petals were developed into flowers; and the other petals remained unchanged. Others in which three were developed; and others where four were developed. In some cases, too, the development of a petal into a flower was imperfect, in the absence of the flower-stalk—the flowers were sessile in the place where the petals would have been. In one case there was an imperfect flower sessile; another imperfect flower on a short stalk; and three perfect flowers on long stalks.
“I was in some doubt whether the petals or the stamens were developed. In cases of imperfect transformation the petals at the base of the umbellule seemed to stand in the position of calyx or involucrum, giving the idea that the stamens were developed into flowers. But in the case where there were six flowers developed there were no petals at the base.
“That it was a matter of extra nutrition was shown by this:—
“1. That they were cultivated as garden flowers.
“2. That where there was one perfectly developed umbellule, it was the only one in the umbel.
“3. That where there were three umbellules they were all imperfect.
“4. That in this imperfect umbellule the perfect flowers were on long stalks and the imperfect ones sessile.