601 
of Edinburgh , Session 1874 - 75 . 
scarcity as adding to the value of a tiling, just as the physical 
dearth adds to the moral dearness of an object ; parco also in Latin 
may be cognate with the English spare. 
The initial l in Latin seems often to disturb the formation of 
words, and to sacrifice some letter that had preceded it. Besides 
other examples already referred to, we may notice the word lien, 
which appears to have lost the older initials sp , which would iden- 
tify it with spleen. But the most remarkable instance of this is 
found in the word lis , which from the old grammarians we know to 
have been originally stlis. But acting on the principle which 
identifies l and r , we see that the original form would give us strits 
instead of stlis, and thus we should have the word commencing 
with the same letters as our Teutonic words strife, sturt, streit, G., 
&c., which undoubtedly are cognate in their meaning with the 
Latin lis, though this word by a strange metamorphosis has lost 
all trace of that struggle or violent contention which it really 
represents. 
It is a peculiarity of the G-reek language that all words beginning 
with p are supposed to have a prefixed aspirate which is analogous 
to a Latin sibilant. If we transferred a Greek word of this kind 
to a Teutonic form, we should prefix a sibilant to the r, but as sr 
is not a combination favoured by the Teutonic languages, a t 
might be inserted for the sake of euphony. Upon this footing we 
may plausibly consider the Greek pevpa as identical with the 
Teutonic stream. It is remarkable, however, that the Gaelic has 
no objection to the initial combination of sr, and accordingly we 
find the word srutli, pronounced srhu, meaning a stream or current, 
and occupying an intermediate position between the Greek pew, 
pew o), and the Teutonic stream. If we could get over the change 
of vowel, we might in the same way connect the Greek pcv, a nose, 
with the Gaelic sron. 
It seems a remarkable circumstance that in Greek and Latin 
words beginning with a sibilant and another consonant there is a 
tendency to confound or corrupt the second consonant so as to 
change one for another in a somewhat arbitrary way. In Latin, 
in particular, where the consonant succeeding the sibilant is 
always a tenuis, one tenuis is frequently changed for another in 
comparing Greek and Latin words. Thus <n rovBrj and studium 
