560 Proceedings of the Royal Irish Academy. 
lu the foregoing cases, where the differences increased from 10^ 
through 13°, 18°, 20°, 22°, the moistened blotting paper occupied the 
upper half of the inverted cylinder; but on removing the moistened 
paper to the lower half, the order of the effects was reversed, the 
secondary index affording the descending series of differences 20°, 17°, 
13°, 12°, 11°, 9°. The glass cylinder was 8 inches high, the hygro- 
meter 5 inches, the cylinder of moist paper 3 inches. These details 
show in how small a volume of air an embarrassing difference of mois- 
ture may exist and derange the results. 
It then occurred to me that, by placing the moist paper in the 
intermediate space between the two situations, the differences might 
make a nearer approach to equality. On making the trial, I obtained 
the following results, viz, : — 
Secondarv, 12° Difference, 12° 
„ ' 24 „ 12 
38 „ 14 
50 „ 12 
62 „ 12 
72 „ 10 
Here the differences may be considered virtually equal (average 
1 2°), no other cause being assignable than the relative collocation of the 
moistened blotting paper ; when this was at the bottom of the glass 
cylinder, the indications of the secondary index diminished from the 
top downwards ; but when at the top, the indications increased from 
the top downwards ; therefore, the most equable indications are ob- 
tainable when the moist paper is placed midway — a hint which may 
be made of more generally useful application. 
Trifling irregularities in the twist of the gut-line occasiop. equiva- 
lent irregularities in the movement of the indexes. When there is 
any doubt or uncertainty about the digit of the decade indicated, the 
following may be employed : — Divide the known length of the portion 
of the gut-line intercepted between the India-rubber disc and the 
clamp-screw, into the whole length of the gut-line ; the quotient is 
the total number of rounds of which the primary index is capable, 
between the extremes of siccity and humidity, which cannot exceed 
ten ; and this quotient, divided into 1 00°, or one round of the secon- 
dary, gives the number of degrees of the secondary that are equal to 
one round of the primary index ; which number, divided into a round 
of the secondary, gives as quotient the number of times which the 
primary index has gone round in the absence of an observer. 
Eetween the results of the two methods there is frequently a dis- 
crepancy, perhaps due to the difficulty of knowing where in the gut- 
line the intercepted portion begins and ends within the disc of India- 
rubber. I have assumed the middle of the thickness of the disc as 
affording a result which cannot be far from the truth. When the 
length of the gut-line (including the intercepted portion) was 3*625 
Primary completed, 1 00° 
200 
300 
400 
500 
600 
