NA TURE 



[May 2, 1907 



into the form of a hollow cylinder by means of the blow- 

 pipe, which cylinder is subsequently opened out into a flat 

 sheet. 



(3) Spun or crown glass, in which a bulb-shaped mass 

 of blown glass is converted into a disc by rotating it rapidly 

 whilst in a plastic state. 



The earliest known window glass, that of the Romans, 

 was produced by the first of these methods ; in fact, it has 

 hitherto been supposed that the Romans knew no other 

 way of making glass into sheets. A careful examination 

 of ' the glass brought to light by the excavations at 

 Silchester, however, leaves no room for doubt that the 

 Romans were acquainted with the art of making blown 

 window glass according to the " muff " process. 



There are reasons for thinking that the making of 

 window glass was not handed down from the Romans, but 

 was re-discovered in the Middle .'Xfies, and the author 

 thinks it most probable (although evidence is too scanty to 

 justify this as a statement of fact) that the glass of the 

 earliest stained-glass windows, that is, those of the ninth 

 and tenth centuries, was made in the same way as the 

 cast window glass most generally employed by the Romans, 

 this being the method that would most naturally suggest 

 itself in the first place. 



The composition of the glass described by Theophilus 

 was quite different from that of the Romans, being pro- 

 duced, according to the treatise referred to, by heating a 

 mixture of sand and the ashes of beechwood. It is prob- 

 •ible, however, that the glass varied very considerably in 

 composition at different times and in different places, 

 owing to the impurity of the sand used and the varying 

 nature of the \vood ashes, which would in all probability 

 be obtained by burning whatever species of timber came 

 nearest to hand. 



Considering these facts, mediseval glass would be ex- 

 pected to reveal on analysis a fairly complicated composi- 

 tion, and, whilst varying considerably, to show in general 

 a richness in alkali (usually potash) and poorness in lime. 



In order to obtain further information on this point, the 

 author, in conjunction with Mr. Percy Williams, deter- 

 mined accurately the composition of some typical speci- 

 mens of medieval stained glass, of which a description is 

 given in the paper. The following table is a summary of 

 the results obtained : — 



C om\iOsition oj Mediaeval Glass. 

 Material Sandiacie Dale Abbey 



54-01 ... 46-94 . 

 4-18 ... 4-II 

 13-20 

 1-70 

 17-37 



S'33 

 2-41 

 0-81 

 1-03 

 0-21 



Mode 



70 



ffla 



16-96 

 0-12 

 19-01 



5 00 

 3-02 

 I -46 

 1-37 



2-l6 



IS 



Silica 

 Phosphoric acid . . 



Potash 



Soda 



Lime 



Magnesia 



Alumina ... 



Iron 



Manganese 



Moisture due to decay 



These results are of interest in several directions, chiefly, 

 however, as revealing the presence of a considerable amount 

 of phosphorus in the glass. The evidence this affords as 

 to the making of the glass, and its effect on the nature 

 of the glass, are discussed in the paper, and attention is 

 directed to the disturbing influence this unlooked-for in- 

 gredient has on the process of analysis. 



The importance of this point lies in the fact that if a 

 partial analysis of the glass is made — with the view of 

 determining the amount of alumina, for example — or if an 

 analysis is attempted without the possibility of the presence 

 of phosphorus being recognised, the results are liable to 

 be seriously inaccurate. A sutnmary of the method adopted 

 by Mr. Williams for obtaining the analyses quoted follows, 

 this being based on the removal of the phosphorus by 

 silver carbonate. 



The paper then passes on to discuss the phenomena of 

 the decay of old stained glass in the light of the peculiar 

 composition revealed by these analyses. 



Medieval glass decays in a very characteristic manner. 

 Very commonly the glass becomes covered with little pits, 

 for all the world like the worm-holes one often sees in an 

 old oak cabinet. 



NO 1957, VOL. 76] 



The process of decay in glass is undoubtedly a parallel 

 on a small scale to the change produced on a large scale 

 by the action of time and weather on geological formations, 

 such as chalk and sandstone — a combination of corrosion 

 and internal change. 



Corrosion of the surface of glass is produced by the long- 

 continued action of moisture, which gradually extracts the 



-Photomicrograph i 



i of deco-Tipo&itic 



val gla 



soluble silicates, leaving the insoluble silica in a thin film, 

 the glass thereby becoming iridescent. Owing to the 

 large proportion of lime it contains, however, mediaeval 

 glass does not become iridescent as the result of corrosion. 

 On the extraction of the alkali by water this lime is left 

 behind with the silica, and forms with it a hard, insoluble 

 silicate of lime, which adheres to the corroded surface of 

 the glass, forming an opaque scum or patina. In some 

 cases this is so marked that the glass appears to be covered 

 with a coat of cement. 



The peculiar pitting of old stained glass is not, how-ever. 



:iograph of similar structures produced experil 



in the author's opinion, due to corrosion at all, but to a 

 change in the constitution of the glass. .As is well known, 

 glass changes its constitution and becomes crystalline if 

 kept at a red heat for a length of time. What happens 

 in a few hours when the glass is hot tends to take place 

 on prolonged exposure to the atmosphere, with this differ- 

 ence, that when the glass is molten its molecules can 



