BY R. GREIG SMITH. 139 



freely like the arabin from Bad. acacia', but seemed to swell like 

 the natural gum of Hakea. The mucilage was tested, and it gave 

 the reactions for arabin, that is to sa}', it was coagulated by 

 alcohol, basic and ammoniacal lead acetates, by ferric chloride. 

 Fehling's solution gave no precipitate, although 1 % copper sul- 

 phate followed by potassium hydrate gave a colourless coagulum. 

 Barium hydrate gave a slight precipitate, while other reagents 

 were inactive. 



When hydrolysed by boiling with 5 % sulphuric acid for five 

 hours, the gum yielded reducing sugars, and these were readily 

 recognised as arabinose and galactose by means of their osazones. 



Thus the organism was a race of Bad. acacioi, midway between 

 that bacterium which produces arabin and its modified form 

 which gives metarabin, and which I have named Bad. tnetara- 

 binum. Although the cultural characters were not typical of 

 either of the forms, yet the points of variance were not suffi- 

 ciently pronounced to say that it was not Bad. inetarahinum. 

 the diagnosis of the gum bacteria is not an easy matter, 

 because the cultural characters of some of the races are extremely 

 variable. Slight alterations in the chemical constitution of the 

 gum influence its degree of solubility, and that controls the 

 microscopical appearance of the colony. The colour is another 

 point upon which too much reliance cannot be placed. That has 

 been emphasised by my researches upon Bac. levaniformans and 

 upon Bac. pseitdarabiiius, both of which occur as white or as 

 yellow races. 



Meanwhile the natural gum of Hakea saliyna had been under 

 investigation to see if it was really a variety of metarabin. The 

 swollen gum was useless for the purpose of testing with reagents, 

 as the drop of mucilage was too gelatinous to mix with the drop 

 of reagent. By subjecting the gelatinous suspension of the 

 swollen gum to a pressure of three atmospheres in the autoclave, 

 a thin solution of a soluble modification was obtained, and upon 

 evaporation this yielded a good mucilage which was tested with 

 several reagents. Basic and ammoniacal lead acetates gave trans- 

 parent clots. Barium hydrate gave a white precipitate, but with 



