Consideration of results
The basis for comparison of the pyrethrin contents depends upon the original samples containing an equal quantity of pyrethrins per unit of dry weight. Neglecting such a sampling error is probably justified to some extent, considering the uniform nature of the bulk sample, and also the total pyrethrins in duplicate samples of undried flowers directly extracted by petrol ether, were found to agree, within the limits of esperimental error, in all three cases. There was, however, very wide variation in moisture content which necessitated estimating the dry weight of each sample separately. This involves considerable inanipulation in the treatment of thc undried samples, but possible error has been minimized as far as possible by determining all known losses of material, such as small quantities left in the grinding mill, on glass rods, and on the filter paper used in the extraction.
The estimation of pyrethrins in the fresh flowers depends largely on the assay of the main extract, containing about 95% of the total pyrethrins. These assays were made in quadruplicate, that is, duplicate assays on each of two samples.
During the extraction of the residue, it is possible that some decomposition may occur during the seven-day maceration, but this is only likely to be slight and should not affect the final result. In the case of the vacuum drying, however, some decomposition certainly occurs. However, it would require a loss of about 50% of pyrethrin I to decrease the final result by 0.01, and that this large loss does not take place is indicated by the methanol extraction, with which the results agree to the nearest single figure. On the other hand, the pyrethrin I1 in the a s e of the methanol extract is considerably greater than that of the extract of the vacuum-dried residue. A solvent such as methanol is likely to give a high value for pyrethrin 11, for instance similar high values have been obtained with acetone and chloroform extracts.G The discrepancy is so great, however, that it may well cover a considerable loss of pyrethrin I1 incurred by the drying treatment.
Comparing the total pyrethrin I of fresh and air-dried flowers, the latter show I-Z% greater values. Apart from experimental error, this is difficult to account for, except by possible defects in the analytical method, caused by material other than pyrethrins formed during drying being estimated as pyrethrin I. The amount is, however, small. The difference between pyrethrin I1 in the fresh and air-dried flowers is much greater, amounting to approximately 8%, 12% and 9% in flowers A, 13 and C, respectively, the fresh flowers giving the larger values in this case. Of the dried flowers, those dried at 55" show slightly less pyrethrins than the air-dried, and the commercially dried considerably less. This is as espected, the lower figures for the commercial dryer being due to prolonged heating for 24 hours. Similar results were found by S. G. Jary and others' using a Iarge-sale dryer.
It may be concluded, therefore, that chemical evaluation indicates that the undried flowers contain about the same amount of pyrethrin I and about 10% more pyrethrin I1 than the same flowers dried by the most favourable natural or artificial means, and 3%-4% more pyrethrin I and x2%-13% more pyrethrin I1 than ffowers dried by one particular commercial method.