Full text: The collected mathematical papers of Arthur Cayley, Sc.D., F.R.S., late sadlerian professor of pure mathematics in the University of Cambridge (Vol. 13)

915] 
ON THE PARTITIONS OF A POLYGON. 
113 
32. To fix the ideas, I assume p = 4, and thus consider the problem of the division 
of the (2w + 2)-gon into quadrangles. Writing 
w — 1 = a+ b + c, 
we take at pleasure any one side of the (2w 4- 2)-gon, making this the first side of 
a quadrangle, the second, third, and fourth sides being diagonals of the polygon such 
that outside the second side we have a polygon of weight a, outside the third side 
a polygon of weight b, and outside the fourth side a polygon of weight c. Any one 
or more of the numbers a, b, c may be = 0; they cannot be each of them = 0 
except in the case w = 1. The meaning is that the corresponding side of the 
quadrangle, instead of being a diagonal, is a side of the (2w + 2)-gon, viz. there is 
no polygon outside such side. Suppose, in general, that P w is the number of ways 
in which a polygon of weight w can be divided in quadrangles, and let each of the 
polygons of weights a, b, c respectively, be divided into quadrangles: the number of 
ways in which this can be done is P a P b P c ; and it is to be noticed that, if for 
instance a= 0, then the number is — P b P c , viz. the formula remains true if only we 
assume P„=l. The number of partitions thus obtained is £P a P b P c , where the 
summation extends to all the partitions of w — 1 into the parts a, b, c (zeros 
admissible and the order of the parts being attended to). And we thus obtain all 
the partitions of the (2w + 2)-gon into w parts; for first, the partitions so obtained 
are all distinct from each other, and next every partition of the (2w -f 2)-gon into w 
parts is a partition in which the selected side of the (2w + 2)-gon is a side of 
some one of the quadrangles. That is, we have 
P w — %P a P b P c (a, b, c as above); 
and it hence appears that, considering the generating function 
f= 1 + P 1 x + P%x 2 + P s x 3 + ..., 
we have 
/= 1 + æ f 3 - 
The reasoning is precisely the same if, instead of a division into quadrangles, we 
have a division into p-gons; the only difference is that instead of the three parts 
a, b, c, we have the p — 1 parts a, b, c, and the equation for / thus is 
/= 1 + x P~ l - 
33. Writing for a moment 
f — U + xfP- 1 , 
and expanding by Lagrange’s theorem, we have 
1.2... w K 
1.2 ... w 
viz. after the differentiation, writing u = 1, we have 
[(p-l)w] w ~ l 
[ w ] W -i 
where it will be recollected that, for the number of sides of the polygon, we have 
r = (p— 2)w + 2. 
In the case of the partition into triangles, p = 3, and we have the before-mentioned 
value 
C. XIII. 
[2w] w_1 + [w\ w ~\ 
w = r — 2. 
15
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.