272 TAXIDERMY AND MODELLING 



intervention of the molecules of glue ; hence, probably, arises 

 the tiresome behaviour of any modelling-glues which contain 

 a greater proportion of oil than Formula 96, some of the models 

 made from them breaking and being unduly affected by changes 

 of temperature. Even this objectionable cracking and break- 

 ing, however, may be obviated in great part by stirring in 

 either dry red lead or flake white or tube oil-colours, but there 

 is no need to make inefficient glue-compositions in face of the 

 admirable behaviour of Formulae 96 and 97. Others which were 

 made contained either too much water, too small a quantity of 

 glycerin, common glue instead of best, or boiled oil instead 

 of linseed, and were unsatisfactory. In none is the weight 

 or expense serious, as i lb. of any of the compositions will 

 model a 2,^ lb. perch, if the model be hollowed out when 

 made, and, when cast and trimmed, the weight will be |- lb. 



To make really good casts with this, free from " blows " 

 and other defects, is by no means an easy matter unless 

 certain rules are observed ; these are — 



1. Brush the plaster mould thinly with warmed colza-oil, 

 and on no account let it be varnished ; then let it be well 

 warmed all through — if small, in an oven, but not in front of 

 an open fire, or the mould will be liable to crack. 



2. Let the modelling-glue be hot, but not boiling ; otherwise 

 bubbles will burst on the surface of the mould and leave 

 holes in the model. 



3. Small objects such as caterpillars, slugs, and snails, 

 minnows, gudgeons, etc., amongst fishes, and small newts and 

 snakes may be cast solid, but larger objects, even fishes so 

 small as \ lb. in weight, as well as larger fishes and 

 reptiles, should be cast hollow ; and there are two very good 

 and sufficient reasons for this, the first being on the score of 

 weight and expense, and the other because, when made thick 

 and solid, the tendency to " blow " is accentuated, and, when 



