The laboratory of the National Canners’ Association is frequently asked to examine samples of tomato products to determine whether or not they comply with the Government requirements. In examining these samples we use the Government method (the Howard method), but do not participate in the discussions regarding its merits and shortcomings. It is our experience that skilled analysts can check themselves and each other with reasonable accuracy, and it is our duty to tell the manufacturer whether his product is legal. Should the Bureau of Chemistry adopt some other method as preferable to the Howard method, it would be our duty to use the new method and continue to serve the industry by telling the manufacturer whether samples submitted by him would pass the Government tests. With a full understanding of our attitude in this matter many manufacturers of tomato products send samples from time to time for examination. It is made plain in every instance that the results obtained by the examination of a particular sample refer only to the batch from which that sample was taken and may give no indication of the character of any other batch. Some manufacturers of tomato products use the Howard method as a check on their factory control. For this purpose it is not satisfactory to have samples examined in a laboratory located at a distance from the factory. Even if several samples are examined from a day’s run, they probably do not represent all the pulp manufactured on that day. It sometimes happens that one wagonload of tomatoes is almost entirely free of rotting material, whereas the succeeding load contains a considerable amount. Even with inefficient sorting, the pulp made from the first load will show a low microscopic count whereas, unless sorting is exceptionally good, the pulp made from the second load may show a high count. Thus one batch may readily comply with the requirements of the Bureau of Chemistry and the next batch may be outside of those limits. Because of this fact this laboratory recommends that manufacturers of tomato pulp do not rely upon the microscopic results of a single sample. The only way in which the product may be absolutely controlled by means of the microscopic count is to examine a sample from each batch—that is, from each kettleful or tankful that is evaporated. This is manifestly impossible. It would require several analysts for one plant. Moreover, it is entirely unnecessary. It has been found that much better results can be secured by having an analyst in the plant to examine samples from time to time. Then, whenever the microscopic count becomes excessive, he can locate the trouble and see that it is corrected.