MEMOIR ON COI-OUREP SHADOWS. 25 



When the number*of the (liadows is three, one of them is of three j 

 alnfoft always complementary to the other two. 



When there are four (liadows, two of them are fometimes of four, 

 complementary to the other two; at other times, one of the 

 ih^dows is complementary to the remaining three. 



When there are only two coloured ihadows, one of them is One of the fha- 

 always of the colour of one of the bodies which refle(6ls the fg^fJJ'e colour as 

 light. If the apartment, in which the fhadows are feen, is the reflected 

 commanded by a covering of flates which reflects a bluilh light "S^'» 

 on the plane where the fliadows fall, the two colours are blue 

 and orange ; if the covering is of tiles, the colours are nearly 

 the fame. If a meadow or trees refled light into the apart- 

 ment, .the ftiadovvs are green and red. If the apartment has which may be 

 coloured hangings which refledl the atmofpheric light or ^^at^^^^^^ j. ^^^"" 

 of the fun, the fliadows will participate of the colours of the 

 lights, reflected externally and internally. 



When the light is refleded through glazed windows upon or by paffing 

 the plane where the fliadow is feen, as the glafs is rarely co- ^'^^Q^S^ S^afs, 

 lourlefs, and generally tinged with green or violet, according 

 as iron or manganefe predominates in its compofition, the 

 fhadow participates in the colour communicated to the light in 

 its paflage through the glafs. 



If the furface on which the dlfl^erent lights are received be or by the colour 



ilfelf coloured, the colours of the ftiadovv will be aflfbaed by it, °^J^f fujfaceoa 



.1 , r .11 1 ,. I which the iha- 



and the problem will become more complicated. dow is received. 



From the obfervations we have collected, it follows na- Inferences. 

 turaliy, that the coloured fliadows of the interior of an apart- 

 ment are capable of yielding all the colours of the fpedrum ; 

 that thefe colours arfe governed by thofe of the bodies which 

 reflefl the light and illuminate the furface on which the co- 

 loured fliadows are obferved. Thus far thefe obfervations Other comple- 

 ofl^er nothing which could not have been forefeen by every ^^^^^^V "^^" 



" J J dows not pro- 



perlon who had refleded a littl^ on fliadows; but that which duced by dired 

 is remarkable is, that thefe fhadows are almofl: always accom- ^^^^^*°^* 

 panied by other fliadows whofe colours are complementary to 

 the refledled colours, although no body is difcovered which is 

 capable of refle6ting thefe forts of colours. 



Here then is a new fliadow which feems conftantly to ac- 

 company the firf^, whofe colour has a peculiar affinity for 

 that of the light which produces the primitive fhadow, and 

 depends on caufes which we purpofe to examine in a fubfe- 

 .quent memoir^ 



Part 



