546 Transactions of the Society. 



Ml' .my fibre. Indeed, in most eas^s by removing the sizing (either 

 rosin or gelatin) and washing the fibres gently with the aid of a 



rather stiff brush, the greater part of a dendrite can be removed 

 from the surface of the paper in which it occurs, leaving in many 

 instances nothing but the metallic nucleus surrounded by fibres 

 stained slightly by a yellowish tinge of colour. The growth of a 

 dendrite is affected to some extent by the nature of the fibres, 

 being most vigorous upon softer cellulose such as that of cotton 

 and certain woods, and least active upon harder cellulose, such as 

 that of linen, esparto, and straw. 



2. The rate of growth of a dendrite in paper is very variable, 

 according to its guidance by variety of circumstances, both internal 

 and external. We have seen that early writers on the subject re- 

 garded the growth as an extremely slow process — as long as twenty 

 years being allowed as an estimate by Tait. From observations 

 of papers, of which the exact date of manufacture was known in 

 each instance, I have ascertained that the oxidation of the metallic 

 nucleus, and consequent growth of the dendritic oxides, is in many 

 cases a comparatively rapid phenomenon ; under favourable cir- 

 cumstances a dendrite may develop appreciably in a period as short 

 as six months. The chief external factor favouring the growth of 

 dendrites in paper is the presence of atmospheric moisture, without 

 which oxidation could not proceed. Cellulose absorbs from 

 9-12 p.c of natural atmospheric moisture, and its moisture-content 

 varies according to the amount of moisture in the atmosphere, and 

 the temperature of the latter. The internal circumstances govern- 

 ing the rate of growth are, the nature of the fibres themselves, and 

 the presence of materials other than cellulose in the fabric of the 

 paper. From numerous observations it appears that dendrites 

 grow most rapidly upon wood and cotton celluloses, less rapidly 

 upon linen, and least rapidly upon esparto and straw fibres. Den- 

 drites grow more readily and rapidly in rosin-sized papers than in 

 papers sized with gelatin (tub-sized), which is due probably to the 

 fact that the gelatin layer on the surface of the latter forms a more 

 complete protection to the fibres, from atmospheric influences, than 

 the rosin-sizing, which latter is chiefly in the interstices of the 

 paper as particles, and to a much less extent as an actual coating 

 upon the fibres themselves. The presence of filling and loading 

 materials such as kaolin and satin -spar is not favourable to the 

 growth of dendrites, and they are consequently rare in heavily- 

 loaded or coated papers. As an example of rapid growth under 

 favourable circumstances, I quote the example of a paper com- 

 posed of wood cellulose 70 p.c, straw cellulose 30 p.c, rosin-sized, 

 10 p.c. loading, which was made in December 1906. Dendrites 

 were first observed in this paper in May 1907, and continued to 

 develop to the present date. The nuclei consisted of small particles 

 of bronze, " 2-1 ■ 5 mm. in diameter, derived from the bars or 



