1900] Leavitt, — Reversions in Berberis and Sagittaria 153 
a type now lost by B. Thundergii, but retained in many species, — the 
bristly-margined type. 
Berberis Thunbergii has by some authorities been treated as a 
variety of B. vulgaris. When we compare the seedlings we seem to 
find very positive evidence against any such assignment. If the two 
forms were so very closely related, their seedlings should be nearly or 
quite indistinguishable ; we see on the contrary that they differ more 
markedly than the mature plants." 
A point to be noted in both forms is that the reverting leaves ex- 
hibit an added character, as compared with the leaves of the adult 
plant. It would be easy to dispose of many cases of suspected pri- 
mordial structure simply as cases of arrested development, in the sense 
of aborted growth. But here the feeble, seedling plant produces organs 
in one respect more highly organized than like organs of the mature 
plant. 
The trifoliolate leaves of B. repens, Figs. 4 and 5, are not the full 
character leaves of that species ; though to judge from a good number of 
herbarium specimens examined they are of rather common occurrence. 
The prevailing leaves have five parts, or even seven. The contour of 
the end leaflet is then altered (Fig. 6). 
If, as is likely, B. vulgaris, B. ZÀhunbergit, and B. Agapatensis are 
really unifoliate, then the trifoliate leaf of B. repens is probably near 
the type from which they have all been derived by reduction. B. 
Aquifolium, a more complex form, reverts only occasionally from its 
more advanced position to the type characterized by the cordate ter- 
minal leaflet; at least in the adult plant. 
SAGITTARIA. In estimating the value of youthful characters for the 
reading of family history, we must take into account the influence of 
the requirements of the seedling, as differing from those of the adult. 
The forms of cotyledons, for instance, are traceable to the form of the 
fruit, the need of storing nourishment, and the exigencies of confinement 
in small spaces. The contours rarely bear any relation to the contours 
of the ordinary foliage leaves, present or past. Similarly after germi- 
nation the young plant may encounter problems that never recur 
subsequently, and meet these problems with specialized structures 
having no connection with ancestral mature types. In the Sagittaria, 
for example (Fig. 7), the interpretation of the linear first leaves (a, 2, c) 
is complicated by the fact that the seedling is submerged, while the 
hastate or sagittate adult leaf is aérial. Was there ever a time when 
