Notes on the Fertilizer Act. 37 



absolutely exact; the error may be infinitesimal, but it is there never- 

 theless. Then again there is the personal factor ; cases arise in which 

 one analyst may consistently get higher results than another analyst 

 using the same process and working- under the same conditions. 

 People are not machines, and no two people are exactly alike. It 

 must be noted, however, that the errors in the analysis may be either 

 plus or minus errors ; that is, if the analyst repeats the analysis many 

 times some of the results will be too low, some too high, while a few 

 may be exact. There is just as much likelihood that the result will 

 be too low as that it will be too high. The analytical error is as a 

 rule very much less than the error in sampling and mixing. The 

 limit is fixed at a figure which it is believed is higher than any 

 ordinary errors of sampling and analysij^ can possibly be. This is 

 in order to give the guarantee every benefit of these errors. We now 

 see that a vendor who deliberately guarantees more of any ingredient 

 than he knows the fertilizer or farm food to contain loses this benefit. 

 Let us take the example quoted above of a bone meal guaranteed to 

 contain 24 per cent, of phosphoric oxide, but actually containing 

 23 per cent. Supposing the analyst to report 22.8 per cent., the bone 

 meal will be reported as not sufficiently in accord with the gii;" 

 composition and the vendor is liable to prosecution. Yet the difference 

 between the 23 per cent, actually pi-esent and the 'liZ.S per cent, found 

 may well be due to errors of sampling and analysis. Had the vendor 

 guaranteed only the 23 per cent, actually present he would be pro- 

 tected against the small error of 0.2 per cent. As it is, he has 

 deliberately thrown away this protection by an over-guarantee. 

 Merchants should take this into account in framing their registrations 

 and guarantees. The registration figures are presumably based on 

 the results obtained by an analyst employed by the vendor. Such a 

 figure is also open to errors of sampling and analysis, and these errors 

 may take the opposite direction to those which come into play in the 

 official analysis. Manufacturers would be well adyised to round off 

 the figures furnished by their analysts. For example, if the analyst 

 reports 15.7 per cent, water soluble phosphoric oxide in a. superphos- 

 phate, the safest plan is to guarantee 15.5 per cent. No monetary loss 

 is incurred, as there can be no appreciable difference in selling price. 

 Many firms adopt this principle, and it is certain that they are not 

 losers thereby, as the security that their guarantees will never be 

 challenged and that a prosecution on this score w\il] never be instituted 

 (and a prosecution means a great deal, even if there be no conviction), 

 will amply pay for any small but doubtful loss sustained by under- 

 guaranteeing. 



Conclusion. 



Two points may be brought to the notice of vendors. Firstly, 

 thait a range of figures, e.g. " nitrogen 2 to 3 per cent." is not 

 admissible in either registration or invoice, and, secondly, that per- 

 centages must be given to one decivial jjlace only. 



These provisions, contained in the footnote to Regulation No. 2, 

 do not seem to have sufhciently attracted the attention of those 

 responsible for the completion of the registration forms. 



It will be seen that the buyer of fertilizers and farm foods is 

 protected by the guarantee from having worthless articles imposed 

 upon him. He can no longer be misled by specious advertisements, 



