340 THE THEORY OF THE STRUCTURE OF PLANTS. 
us to extend the principle to this latter, and I cannot but conceive that 
a more probable explanation has been proposed. 
My note on the structure of Primulaceze relates to one point which 
I have not seen rightly explained. In this order the stamens are 
observed to be constantly opposite the petals, a circumstance which 
always seems to need some explanation. In the present case, I think 
it evident that it is due to the abortion of a circle of parts belonging 
to the intermediate position between the petals and stamens and 
- alternating with both. A eareful examination of almost any Primula, 
the Auricula affording an excellent example, shows that the coloured 
eye of the flower consists of a series of pieces like the petals, as it 
were fastened on to them, and in such an order that the middle of 
each arch of the eye is exactly placed between two of the petals. In 
the genus dretia this is still more evident. In Samolus a set of abortive 
stamens occuis between the petals, and the same is the case with 
several species of Lysimachia; in Cyclamen this organ is also easily 
observed, and in Glaux the proper corolla as well as its double is sup- 
pressed. From these examples we are enabled ideally to restore the lost 
circle, where it is most completely suppressed, and thus to compre- 
hend the true symmetry and the reason of a seemmg departure from a 
general rule. In how many other cases of opposite circles a similar 
explanation may be justified, I will not presume to say. In respect, 
to this order I think it entirely satisfactory, but it is not the only one 
conceivable, for any one who has carefully considered a Camelia, in 
which the numerous circles of petals, instead of alternating as is usual, 
are forced into regular lines radiating from the centre, will be ready 
to admit the possibility of parts which are normally alternate becoming 
opposite by a sort of twist, and what occurs occasionally as a variety, 
may occur uniformly or nearly so, from a like cause, more constantly 
operating on a particular tribe, so that we are by no means driven to 
imagine without evidence an intermediate circle, in every instance of 
opposite parts, nor is there any necessity for assuming the occurrence 
of Chorisis where it cannot be distinctly proved. 
