Donovan — On the Comparable Self-acting Hygrometer. 561 
inches, and the intercepted portion -56 inch, dividing the latter into 
the former, the quotient 6-47 is the number of rounds which the pri- 
mary index should give, by calculation ; but, by experiment, it gave 
seven complete rounds. The discrepancy is easily explained when we 
recollect that of two strings of any kind, the same in all respects, if 
one be moistened, it becomes shorter. As long as the length of the 
gut-line and the ratio of the intercept remain unchanged, the quotient 
of the division of the former by the latter may be used as a check on 
the indications of the two indexes ; and being the natural unit, should 
measure the graduated circle. 
From a number of experiments I give the results of a few to show 
how nearly the secondary index agreed with the results of calculation. 
In nine cases they agreed exactly; in three the difference was 1°; in 
four the difference was 2° ; in three it was 4° ; in three it was 5° ; in 
six it was 6°. These differences are all explicable by the fact that in the 
experimental cases, moisture was concerned, and not in the calculated. 
Eut in a journal where averages for the day or week are to be 
noted, such differences would all but disappear. 
On making trial of a new gut-line to discover how much water it 
contained in what may be called its natural state (i. e., as procured 
from the music seller), I found that when confined in the receiver of 
the hygrometer with an exsiccating disc, the index went round twice ; 
hence the necessity of exsiccation when a new gut-line is to be used 
for an important experiment. 
The means of recording the maximum of moisture in the absence of 
an observer with this instrument are very simple. "Wrap a piece of 
very thin soft iron wire round a common brass pin of the same thick- 
ness as the gut-line, in the form of a helix, consisting of four or five 
coils, so as to form a kind of hollow cylinder of wire, through which 
the gut-line is to pass, and to constitute an axis for the helix to turn 
on. The redundant wire at one end is to be cut off ; and the redun- 
dant wire at the other end is to be bent away from the helix as a hori- 
zontal arm at a right angle with the gut-line, and again bent downwards 
at a right angle with the arm in such a part as will cause it to be 
encountered and carried forward by the secondary index when moving 
in a forward direction. But when the secondary retrocedes in con- 
sequence of drought, it leaves the iron wire index stationary to point 
to the maximum which it had reached during the absence of an ob- 
server. 
