The Cambodia cotton flowers (G. hirsutum) contain a good amount of pigment (3%) in spite of their feeble ivory colour. In one sample the pigment was found to be mainly Quercimeritrin along with a small quantity of Quercetin, whereas in another sample collected next year Quercetin formed 75% of the total pigment and Quercimeritrin only 25%. In the nature of the pigments these flowers differ from the yellow American upland cotton (G. hirsutum) which have been reported to contain Quercimeritrin and Isoquercitrin. It is therefore concluded that the nature of the pigments vary with (1) the variety, (2) the locality, and (3) the season.

The flower petals of the Uppam cotton plant (G. herbaceum) contain as their main components gossypitrin and quercetin. An appreciable quantity of a new glycosidic pigment and a small quantity of gossypetin were also isolated. Uppam petals, therefore, differ in their composition from those ofG. herbaceum examined by Perkin who probably obtained his supply of the petals from the North Indian variety. His results indicated that gossypitrin was the main component and that isoquercitrin was present in small quantities.

A preliminary examination, using a small sample of the flower petals of the Karunganni (G. indicum) collected in 1932, showed that the pigments consisted mainly of Quercimeritrin and Gossypitrin. From a study of a larger sample collected in 1933 it was found that the following were the main components: Gossypin, a complex glucoside of Gossypetin; Quercetin, a new non-glycosidic pigment having the probable formula C₁₆H₁₂O₇ Herbacetin, mostly as its glucoside Herbacitrin, first isolated from Uppam (G. herbaceum) as a new member of the flavonol series was also found to be present.

Herbacitrin, the new flavonol glucoside obtained fromGossypium herbaceum (Uppam) has the formula C₂₁H₂₀O₁₂. Though isomeric with Quercimeritrin it differs from it in many properties. On hydrolysis it gives glucose and Herbacetin having the formula C₁₅H₁₀O₇ which differs from Quercetin in its reactions. In their colour reactions with alkaline buffer solutions and withp-benzoquinone Herbacitrin exhibits a very close resemblance to Gossypitrin and Herbacetin to Gossypetin. From these and other considerations Herbacetin has been given the constitution 3 ∶ 5 ∶ 7 ∶ 8 ∶ 4′-pentahydroxy-flavone and Herbacitrin is represented as its 7-glucoside. This has been confirmed by the oxidation of Herbacitrin top-hydroxy-benzoic acid. The flavonol has now been synthesised in the Dyson Perrins Laboratories in Oxford and found to be identical with the natural specimen.

An attempt has been made to fix the position of the glucose residue in gossypitrin by methylation with diazomethane and hydrolysis of the product. It has been found that methylation is incomplete, one hydroxyl group remaining unaffected, and thereby giving after hydrolysis a tetramethyl-gossypetin. This compound probably has the two hydroxyl groups in the 5 and 7 positions.

It has been shown that hydrochloric acid can be used for dissolving lead oxalate precipitates in the estimation of lead and also for the titration of the liberated oxalic acid with permanganate provided the concentration of the hydrochloric acid is not greater than 0·5 N. Employment of hydrochloric acid obviates the difficulties due to occlusion of oxalic acid in Low’s method and the complications in the permanganate titration arising out of the use of nitric acid. Quantitative precipitation of lead as oxalate does not require 60% concentration of acetic acid. Precipitation is complete even when 5–6% of this acid is present. The use of nitric acid in oxalic acid-permanganate titrations has been shown to be not permissible.

The lanthanum nitrate test for the acetate ion has been examined for use in the routine analysis of mixtures of inorganic substances including the acetate, oxalate and tartrate radicles. A distillation method for the preparation of the test solution and a modification of this procedure in the presence of sulphites and thiosulphates are described. The results obtained are quite satisfactory and the test is recommended for general use.

The colouring matter of the flower petals ofHibiscus cannabinus consists mostly of the glucoside cannabiscitrin along with small quantities of the corresponding aglucone cannabiscetin. The properties and reactions of these two substances are described along with those of some of their derivatives.

Cannabiscetin forms a hexaacetate and a hexamethyl ether. It is a flavonol exhibiting similarities with gossypetin and herbacetin particularly in regard to alkali colour reactions and gossypetone reaction. It yields gallic acid on being subjected to alkali fission, its methyl ether producing trimethyl gallic acid. It is, therefore, concluded to be 3:5:8:3′:4′:5′ hexahydroxy flavone.

Addition of boric acid to O-hydroxy-carbonyl compounds dissolved in concentrated sulphuric acid, except in a few cases, brings about either a marked intensification or a change in colour of the fluorescence exhibited by them in daylight or under the lamp. In a few other cases fluorescence appears for the first time on the addition of boric acid. Since even the simpler O-hydroxy-carbonyl compounds yield these results with boric acid, it appears reasonable to conclude that this reaction will serve as a suitable method for the detection of O-hydroxy-carbonyl group. It must, however, be borne in mind that while the appearance of fluorescence, its intensification or a change in colour indicates the presence of a hydroxyl group ortho to the carbonyl, the non-appearance of the fluorescence effects described does not indicate the absence of this group.

Further work along these lines is in progress.

A new fluorescence reaction for the detection of boric acid with resacetophenone and phosphoric acid has been described. None of the metallic radicles examined interferes while among the acid radicles chromate, bromate, chlorate, nitrite and fluoride do so to a limited extent. The reaction compares favourably with those described in the literature both as regards limit of identification and also interference. It has been shown that the test is applicable for the detection of micro-quantities of boric acid.

A sample of the flower-petals ofThespasia populnea is found to contain populnin (0·33%), populnetin (0·07%) and herbacetin (mostly as its glucoside, 0·03%). The proportions vary with different samples depending on seasonal and other factors. Besides these three yellow pigments, a colourless phenolic compound called populneol is also present (0·16%).

Populnetin seems to be a new tetrahydroxy flavone and populnin is its monoglucoside.

A new method for the approximate colorimetric estimation of boric acid depending on the yellow colouration yielded by boric acid with pentamethyl-quercetin in presence of citric acid in anhydrous acetone medium has been described. The method is empirical in character and gives results with a maximum error of about 15% with quantities of boric acid ranging from 0·4 to 1·8 mg.

Addition of boric acid to aromatic compounds (C, H, and O only) of various types containing the, dissolved in concentrated sulphuric acid generally produces a marked intensification or a change in colour of the fluorescence exhibited by them in daylight or under the lamp In several cases the solution of the compound in sulphuric acid is itself non-fluorescent and fluorescence appears on adding boric acid

Further work is in progress.

The introduction of bromine atoms into the molecule of anO-hydroxycarbonyl compound does not invariably result in absence of fluorescence effects on the addition of boric acid.

The fluorescence test for theO-hydroxy-carbonyl group is also applicable in the case of brominated compounds but as with the bromine-free compounds a negative result does not necessarily mean absence ofO-hydroxycarbonyl group. Negative results are obtained more often when bromine is present than in its absence.

The introduction of the azo-group into the molecule of ano-hydroxycarbonyl compound does not invariably produce a negative result in the fluorescence test with boric acid carried out under filtered U.V. light. Fluorescence intensification on the addition of boric acid is observed in several cases, though there is a marked diminution in the intensity of the effect when compared with the parento-hydroxy-carbonyl molecule.

The fluorescence test for theortho-orientation of the hydroxyl and carbonyl groups is also applicable to the azo-dyes but a negative result does not prove the contrary.

The magnesia and residual manganese have been precipitated together with oxine, the content of the latter determined colorimetrically and that of the former calculated by difference in the analyses of samples of charnockite, leptynite and garnet. The results obtained are compared with those obtained in parallel determinations carried out by the pyrophosphate method and improved accuracy in the estimation of residual manganese as well as magnesia claimed.

Sixty samples of samarskite were collected from a mica pegmatite vein in Nellore district, S. India.

A typical sample of samarskite has been analysed and the results reported.

An upper limit of identification exists for the lanthanum nitrate test for acetate.

Precipitates of manganese oxyquinolinate obtained from a mineral acid solution containing a slight excess of oxine by adding dilute ammonia, according to Berg’s second method, could be dried to constant weight at 150° C. and weighed. A slight superficial discoloration which may be observed in some cases is inappreciable even when dealing with small amounts of manganese. The composition of the dried precipitate corresponds to Mn (C₉H₆ON)₂ and contains 16·03% manganese.

Satisfactory results are obtained in a single precipitation provided a large excess of oxine is not added. A large excess of oxine with or without the addition of strong ammonia tends to yield higher results. Double precipitation does not improve the accuracy of the results in such cases.

The presence of large amounts of ammonium or sodium chloride does not appreciably affect the accuracy of the results.

Small amounts of manganese ranging from 0·3 to 4·0 mg. could be satisfactorily determined in presence of large amounts of sodium chloride and strong ammonia.

Quantitative studies on the bromination of resacetophenone showed that the reaction was slow and progressive and its magnitude depended primarily on the amount of bromate initially added in excess of the theoretical equivalent so that reproducible results could not be obtained directly by Koppeschaar’s method. However, on plotting the volume of bromate initially added against the excess found by titration, a linear graph was obtained which on extrapolation to zero excess gave the value of bromate equivalent to di-bromination of the ketone present in solution. The estimated maximum error involved was of the order of 5%.

The volumetric determination of manganese with 8-hydroxy-quinoline was investigated. Errors due to adsorption have been shown to increase with increase in (a) amount of excess reagent and (b) strong ammonia; and to decrease in presence of sodium chloride. The direct titration method was found to yield satisfactory results provided a large excess of oxine or ammonia was not used. The back titration method was found to yield very high values due to adsorption.