49 8 Agnes Arber.—On the Structure of the A ndroecium in 
mesarch strands in the sucking cotyledons of the Scitamineae can be fully 
explained when their relation to the cortical vascular system is understood. 
The mid-rib bundle of the cotyledon of Persoonia lanceolate described by 
T. G. Hill and E. de Fraine, 1 furnishes another example. Here the bundle 
is normally formed, but is accompanied by extra xylem elements, which 
sometimes give it almost a mesarch appearance. The authors, however, 
interpret these elements as transfusion tracheides, and regard them as 
adaptations to supply the physiological needs of a xerophytic seedling. 
Turning again to the case of Parnassia , we find ourselves confronted 
with the question whether the mesarch structure in this case is merely 
adaptive as in Persoonia. I am inclined to think that this alternative may 
be dismissed, since there seems, in this instance, to be no sufficient reason 
for the development de novo of such an anomaly. 
We have rejected the idea that the mesarch xylem of Parnassia is 
reminiscent of a pre-Angiospermic ancestor. Nevertheless, it is not im- 
probable, on general grounds, that this structure possesses some phylogenetic 
significance in connexion with the more modern ancestry of the genus. 
Vascular strands have a strong tendency to form, as it were, an internal 
record of ancestral features, even when these have become almost entirely 
obliterated as far as external form is concerned. When evolution pro- 
ceeds in the direction of reduction, vascular structure generally, though 
not always, lags behind the outward form, and becomes diminished to 
vanishing point at a less rapid rate than the surface features. In other 
words, when an organ is becoming rudimentary and is on the point of 
disappearing, the branches of vascular tissue which formerly supplied it 
are generally still traceable, though they may be reduced to mere stumps. 
This idea is present, though it is not definitely formulated, in the writings 
of Robert Brown 2 as early as 1833. Brown observed vessels in the 
‘ auriculae ’ of certain Orchids, which were regarded on other grounds as 
representing abortive members of the androecium. These members were 
no longer recognizable as stamens externally, but the vascular bundles, 
which were intended to supply them, retained their original course, although 
their function was a thing of the past. 
A particularly interesting instance of the survival of vestigial bundles, 
when the organ, which they are intended to supply, has become almost 
entirely obliterated, has been described by E. M. Berridge 3 in the case of 
1 Hill, T. G., and de Fraine, E. : On the Influence of the Structure of the Adult Plant upon 
the Seedling. New Phyt., vol. xi, 1912, p. 319. 
3 Brown, R. : On the Organs and Mode of Fecundation in Orchideae and Asclepiadeae. Trans. 
Linn. Soc. Lond., vol. xvi, 1833, pp. 697, 698. 
3 Berridge, E. M. : The Structure of the Female Strobilus in Gnetum Gnemon. Ann. Bot., 
vol. xxvi, 1912, p. 990. (It should be noted that Miss Berridge does not herself accept the 
generalization that, in the process of reduction of an organ, traces of its vascular system survive, 
even at the stage at which its external features are practically obliterated. She writes, ‘ In the process 
of reduction the vascular bundles seem usually to dwindle and disappear before the organ itself is lost, 
