A. Ritchie-Scott 
425 
If now we denote the moments of the quadrant as measured from the 
bounding planes by q^^, C[o\, Q20, etc., with the standard deviations Sx, Sy of the 
quadrant as units, it can easily be shown that 
0 _ ^20 _ W20 - + _ A. {hHA + r (1 - r-) x + r {2h - rk) KB\ + 1 + A- 
^ _ 9o2 _ -"^02 - 2kmn+Jf _ ^ [kKB + r (1 - r^) ^ + r (2A; - r/i) iT^ ) + 1 + k- 
^ _ _ "hi — ^'™oi — ^"^10 + /i^' _ ''^ \hKB + ( 1 — X + /<^-^^-4 } + + A/- 
~ " (m,o-/i)(woi-^) ~ {kiJSA + riTi?) + Aj [\ {rHA + A"5) + k} ' 
$20) Q02. and are known but so far as I can see at present the solution 
of the above equations would require the construction of four-entry tables in 
h, k, r, and X,. 
If these were found, then 
q-2oS/ = {mo,, - 2A?/ii„ + h") 0-^2, 
qnSxS,, - {niu - /""oi - kr7i,n + hk) <jx<J,, 
and 
y 
- \ {rHA + KB) ^ k 
N 
m 
In conclusion my thanks are due to Professor Pearson for suggesting the 
subject of the above enquiry and the interest he has taken in its course, and 
again to Miss Alison Robertson for assistance in correcting the proofs. 
