Parnassia and its bearing on the Affinities of the Genus . 497 
tissue of the filament is relatively reduced. The small bundle contains, as 
a rule, about a dozen xylem elements arranged in a compact circular group. 
There is no ring-structure, but we can sometimes detect the occurrence of 
smaller elements in the centre of the group. If, as seems probable, these 
elements are protoxylem, the bundle may be regarded as mesarch — corre- 
sponding, that is, to that of P. palustris , but in a reduced form. 
Parnassia parviflora is still further reduced. The small filament con- 
tains a little group of xylem elements (sometimes about eight), but the 
protoxylem cannot be distinguished. 
(iii) The Structure of the Nectaries in Parnassia palustris, L. 
Drude 1 has shown that, if a section be cut across the leaf-like expansion 
of a nectary of Parnassia palustris , it is seen to be traversed by a number 
of separate bundles, each of which is destined to pass into one of the 
glandular branches. I have found that these bundles retain their distinct- 
ness down to the level at which the staminode fuses with the receptacle 
(Text-fig. 3, f). As they pass inwards they unite into a single bundle, whose 
origin has been already discussed. Little detail can be made out in the 
vascular structure of the nectary in this species, since the bundles are 
practically unlignified. Eichinger 2 has shown, however, that these unligni- 
fied strands may safely be interpreted as vascular bundles, since, in 
P. viridiflora , Batal., vessels occur in the strands which are directed 
towards the lobes of the nectary. 
III. A Suggested Interpretation of the Mesarch Structure 
in the Stamen-filaments of Parnassia. 
The occurrence of centripetal xylem in connexion with the single 
vascular bundle which traverses each filament in Parnassia palustris and 
P. fimbriata is a peculiarity which seems to call for some explanation, 
since it diverges remarkably from the simple collateral structure which is 
usually to be found in stamen bundles. In general, centripetal wood may 
be regarded as an essentially cryptogamic character , 3 which has scarcely 
survived at all among the Phanerogams. It seems manifestly absurd, 
however, to regard the isolated case of its occurrence in Parnassia as 
a vestigial trait, recalling some remote cryptogamic or gymnospermous 
ancestor. 
The occurrence of mesarch vascular strands is extremely rare among 
the Higher Plants, and it is probable that there is no general explanation 
for the phenomenon, but that each case must be dealt with separately, on 
its own merits. E. M. Berridge has shown , 4 for instance, that certain curious, 
1 Drude, O. : 1. c., p. 260. 2 Eichinger, A. : 1. c., pp. 308 and 309. 
3 Scott, D. H. : The Old Wood and the New. New Phyt., vol. i, 1902, p. 25. 
4 Berridge, E. M. : Note on the Mesarch Structure of certain Vascular Bundles in the Cotyledons 
of some Scitamineae. Ann. Bot., vol. xxiv, 1910, p. 485. 
