OFFICIAL INSPECTIONS 80. 193 



carry 46.20 per cent of lime (calcium oxide) and 12.89 per cent 

 of insoluble matter. 



No. 3729, R-R Ground Limestone, made by the Roekland- 

 Rockport Lime Co., Rockland, Me., was sampled at Kendall & 

 Whitney's, Portland. The sample as received (No. 3729) car- 

 ried the label and the guarantees of ground limestone. The 

 goods, however, were not ground limestone but gypsum. The 

 sample contains 22.40 per cent of lime (calcium oxide), equiva- 

 lent to 68.7 per cent gypsum. The insoluble matter was 4.15 

 per cent. This sample was probably wrongly marked by the 

 inspector, for, so far as is known at this Station, the Rockland- 

 Rockport Lime Company did not register or offer gypsum for 

 sale in 1916. 



No. 4167, Tobique Land Plaster, made by John Stewart, Plas- 

 ter Rock, N. B., was guaranteed to carry 95 per cent of calcium 

 sulphate (gypsum). The sample was obtained from P. H. Reed, 

 Fort Fairfield, and was found to carry 67.06 per cent of gypsum 

 (calcium sulphate) and 24 per cent of insoluble matter. A 

 rather impure, poor quality gypsum (land plaster). ■ 



No. 4294, U. S. Gypsum Plaster, made by the U. S. Gypsum 

 Co., New York City, was guaranteed to carry 73.87 per cent of 

 calcium sulphate. The sample was taken from Sweetser & Cole, 

 Gray, and the inspector stated that he was not sure that it i? 

 the U. S. Gypsum brand. Exceedingly good goods, carrying 

 99.32 per cent of gypsum (calcium sulphate) and 0.85 per cent 

 of insoluble matter. 



Discussion of Rf,sui/ts of Analyses. 



The errors due to the sampling and laboratory examination 

 effect, of course, the results of the analyses as reported. The 

 Station has no control over taking the samples. It rechecks all 

 analyses to reduce the laboratory errors to a minimum. Errors 

 in sampling, due to imperfect methods, improper storage of 

 goods, or other causes cannot be checked up. The larger the 

 number of samples of 'any one brand examined the better is 

 the measure of the equality of the goods as a whole. But as 

 goods are made in comparatively small batches and from more 

 or less different materials it may be that individual lots of 

 brands showing well on the average may differ materially in 

 composition. 

 5 



