470 
Saunders.—A Reversionary Character 
genus is always regarded as dimerous. Prantl and Kundig 1 describe it 
as composed of two carpels with 2-4 stigmas, and the fruit as dehiscing 
either septicidally or by the two valves becoming detached from the placentae. 
Figures are given of an ovary of four segments, with two pairs (one longer 
and one shorter) of filiform stigmas, and of a ripe, split, two-valved fruit with 
a row of seeds on one side only of each valve, both from E. californicus 
(Figs. 82 B and 83 c); also one (from Baillon 2 ) (Fig. 46) of a fruit of 
E. crocea showing two valves with triple stigmas, the valves being detached 
from the two placentae, each of which carries a single stigma. Finally, in 
addition to two filiform stigmas, two others of 5-lobed plate-like form are 
shown by Payer 3 4 in a drawing of E. crocea (Fig. 47). On the accepted 
formula of G 2 there are several points in these statements and figures which 
require to be explained away, viz. the varying number of stigmas (2,4, 8,12) ; 
the two modes of dehiscence (the single split into two seed-bearing valves, 
and the double split into two seedless valves and two seed-bearing placentae, 
with the associated difference in length of the two pairs of stigmas); the 
numerous confused rows of ovules on the placentae. In order to arrive 
at a true interpretation of these somewhat contradictory relations, we must 
take into consideration the following facts: 
1. All botanists to-day are agreed that Eschscholzia is closely related 
to the two genera Dendromecon and Hunnemannia , and that these three 
forms stand very near to a group in which the carpels are more than two, 
viz. Platystigma with three and Romneya and Platystemon with six to ten 
or more. 
2. Eschscholzia 4 (Fig. 48), Dendromecon , and Hunnemania are all 
distinguished by having fruits with ten conspicuous ribs. In the intervening 
flat sections of the roughly no-sided ovary there is an additional subsidiary 
vascular strand. 
3- Dehiscence of the fruit in Dendromecon and Hunnemannia is by the 
two valves becoming detached from the two thread-like placentae, which 
form a frame (replum) as in certain of the Capparidaceae. 
4. The two stigmas of Dendromecon , though stated to have three 
lappets (‘ Pflanzenfamilien ’, iii, 2, p. 138), are shown in the drawing (loc. cit. 
Fig. 87 c, aho ‘ Bot. Mag.’, t. 5134) to be in reality 5-lobed ; 5 that of Hunne - 
1 Pflanzenfamilien, iii, 2, pp. 138, 139. 
2 Nat. Hist., iii, p. 118, Fig. 140. 
3 Organogenie, Tab. XLV, Fig. 38. 
4 See Bot. Mag., 56 (1829), t. 2887. 
On the ground that he himself had never found 5-lobed stigmas in the plants which he 
examined, tedde asserts that these drawings must be incorrect (Pflanzenreich, iv, 104, p. 39). But 
Dendromecon may well show transition stages like Eschscholzia. Not only is there no ground for 
assuming these figures to be incorrectly drawn, but in the light of what has been made clear in the 
present account, such a 5-lobed outline to each valve stigma might well be expected from the outline 
of the ovary. 
