40 6 Mr. de L vc on 
treme drynefs\ which, however, till now, has created no 
doubt. 
85. That lingular thread illuftrates alfo another point, clofely 
connected with the retrogradation of its tribe, but not to be 
confounded with it; I mean the recoil. The common caufe of 
both phenomena is, two oppofite effeSls produced by changes 
of moijlure , on the length of threads . The retrogradaiion in 
'the march , comparatively to that of moijture , is produced by 
•one of the effects, which before was furpaffed by the other, 
becoming predominant ; and the recoil , or a returning back a 
part of a firft ftride when moijture changes fuddenly, is pro- 
duced, by one of the effects, that on the fibres themfelves, being 
performed fooner than that on the fort of reticle formed by the 
fibres . Now, the box thread having its retrogradation at ap- 
proaching extreme drynefs , there alfo the recoil becomes fenfi- 
ble : it appears in the firft modifications of that thread when 
placed in my dry vejfel , by motions backwards and forwards, 
as it happens to fome other threads , when taken out of water , 
or expofed to any other fudden change of moijlure. The box 
thread having a flow motion, no recoil is clearly diftinguifhable 
in its common march ; whereas the hair and the quill- thready 
which, in appearance, are very quick, have generally a very 
difturbed motion when moijlure changes fuddenly. I have feen 
them, when in a free air, and happening to be in their fta- 
tionary ftate, moving quickly one way in a fpace of 1 or 2 de- 
grees, and then recoiling flower, fometimes to the fame point 
where they were before, while my hygrometer underwent a fteady 
change, which was in the firft direction of theirs. This phe- 
nomenon, of a complete recoil in the Jlationary ftate of quick 
threads , is firnilar to the recoil of the index in thofe glafs frames 
that I have defcribed in § 59. of this Paper, which have a 
com- 
