204 Scientific Proceedings, Roijal Dublin Society. 



Calculated on the dry weights of the flowers used, the i'oregoiiig analyses give 

 approximately 140 parts of iron per million for the blue flowers' and 60 p.p.m. 

 for the pink, when the difference in ash-content is taken into consideration. 

 The analyses of Maquenne and Cerighelli (1921) show that the iron in plant 

 tissues varies from about 20 to 150, rising exceptionally to much higher values, 

 such as 362 p.p.m. ; thus the quantities found for hydrangea flowers are quite 

 normal. These workers draw attention to the fact that iron accumulates in 

 tissues such as bark, leaves, etc., as they grow old. Such a difference, however, 

 can hardly be held to explain the varying iron-content of two flowers in full 

 bloom. Moreover, were this the cause, flowers which open pink should later on 

 become blue, but such a change does not occur — the pink remain pink and the 

 blue remain blue, though the intensity of colour may change somewhat. Further- 

 more, the evidence of the hajmatoxylin test is against it, as young blue flowers 

 often give the reaction intensely. 



Correlation of Results with those of Previous Workers. 



It had long been recognised that certain soils possessed the property of 

 producing a blue colour in the hydrangea; and Charles Darwin recorded that 

 alum influenced the flower colour. 



Molisch grew no less than four hundred hydrangeas in various soil mixtures 

 with or without added salts and metallic iron and iron oxides. Since the 

 flowers were normally all red, it is evident that his standard earth must have 

 been alkaline. The addition of iron and iron oxides failed to induce any 

 alteration in colour. This is in keeping witli the results for alkaline soils. In 

 peaty soils the plants grew excellenth'', and produced blue flowers. Various 

 metallic salts were tried, and either had no effect or were poisonous to the plant. 

 Ferrous sulphate, however, led to the production of blue flowers, as did also 

 alum and aluminium sulphate. The action of these salts is to increase the 

 acidity of the soil, for quite dilute ferrous suliDhate solution may be as acid as 

 pH 48. Pure aluminium sulphate in dilute solution gives an acidity of about 

 pH 4, the value for alum being almost identical, as previously mentioned. 



Since iron in some form is always present in the soil, an increase in soil- 

 acidity renders more iron available for the plant, so the addition of alum or 

 aluminium sul^jhate increases both the soluble iron and aluminium salts. To 

 which of these, then, is due the production of the blue flowers? As already 

 mentioned, ferric salts are precipitated as h.ydroxide between pH 35 and pH 60, 

 ferrous from about pH 51 onwards to about pH 8; aluminium, as h.ydroxide, 

 begins to be precipitated at pH 3-9 to 42, a precipitate is still obtained at 

 pH 54, but on filtering the solution, the filtrate at pH 64 fails to give any 

 further precipitate or trace of turbidity when rendered less acid. The quantity 

 of aluminium salts in solution at pH 6 to 64 must therefore be very minute, 

 though hydrangeas are normally blue at pH 6, tending towards mixed colours 

 at pH 6-4. These considerations, taken in conjunction with the qualitative and 

 quantitative tests for iron in the flower, render it probable that iron rather than 

 aluminium is the metal that reacts with the pink anthocyanin to give the blue 

 in the growing plant. An experiment performed by Molisch certainly does not 

 leave this conclusion entirely free from doubt. Having failed, as did also the 

 writer, to obtain an extract of the pigment, \vhich quickly becomes decolorized, 

 Molisch found that longitudinal sections of the flower stalks gave a blue colour 

 when treated with alum, aluminium, sulphate, or ferrous sulphate. 



Since indicator changes in the pigment have been ruled out, the following 

 possibilities present themselves as interpretations. Firstly, tliat under the 



