1 56 Johnson. — On A rceuthobium Oxycedri. 
at the base, and in these cases the anther returns to the usual 
form of this organ. Also, it sometimes (exceptionally) happens 
in those genera that there are flowers with three perianth- 
leaves and only two stamen-leaves, in which case one of the 
latter is placed in the space between two perianth-leaves ; 
certainly the best evidence against Hofmeister’s determination. 
The superposition of stamens and perianth-leaves can be ex- 
plained as in Loranthus V In Arceuthobium the stamen is 
distinct from the perianth-leaf at first, just as it is practically 
throughout life in Eremolepis. The absence of a vascular 
bundle in the stamen finds its counterpart in the evascular 
character of the carpels, which, except for this absence of 
bundles, have all the characteristics of the carpels of a normal 
Angiosperm. It is only by a great strain of comparison that 
the stamens and carpels can be regarded respectively as ligules 
of the perianth-leaves, or as similar to the integument of the 
ovule in Coniferae — hypotheses which have been advanced 
as favouring an affinity of Loranthaceae with Gymnosperms. 
The comparison of the young and old male flowers of Arceu- 
thobium furnishes one more illustration of the representation 
of phylogeny in ontogeny : Arceuthobium , one of the most 
highly modified of the Visceae, passes through a stage in the 
development of its male flower which is permanently repre- 
sented in less modified members of the group 1 2 (e. g. Eremolepis ). 
Allowing for the decrease in the number of layers composing 
the wall of the pollen-sac, the course of development is normal 3 
(Fig. 16). There was however an interval of twelve months 
between the young and old male flowers examined. There 
were no flowers showing stages intermediate between the 
archesporial cells and the nearly ripe pollen-grains. 
6. Vegetative Organs. 
The detailed and fully illustrated description of the vegeta- 
1 Eichler, Bliithendiagramme, p. 554. 
2 Bentham and Hooker, op. cit. on page 138, p. 206. 
3 Goebel, Outlines of Classif. and Sp. Morphol., p. 362. 
