July IS, 191S Methods of Bacterial Analyses of Air 345 



In Petri's (21) first sand filter much coarser sand than that now employed 

 was used and in deeper layers. Two 2.5 cm. layers of a sand which had 

 been heated to redness and which would pass through a 0.5 mm. sieve 

 were supported on the inside of a glass tube on wire-gauze disks. After 

 aspirating a large volume of air through the filter tube by means of an 

 air pump, the sand of each layer (the second acting as a control) was 

 divided between a number of plates and mixed with nutrient gelatin. 

 The number of colonies which developed after incubation was then 

 counted and the number of bacteria originally present calculated from 

 this. 



Soper (26, 27) seems to have been the first to use two sand filter tubes 

 in tandem. He washed the sand in sterile water and made his plates 

 from this wash water instead of adding the sand directly to the Petri 

 plates, as former investigators had done. In this way he obtained more 

 transparent plates and probably a more uniform and thorough admixture 

 than had his predecessors. 



Winslow (33) compared the results from sand filters made of as coarse 

 sand as that used by Petri with those from filters made of sand of o. i to 

 0.3 mm. in size, with much better results for the latter even when the 

 layer of sand was reduced to 2.5 cm. as compared with 5.0 cm. of the 

 coarser sand. 



Weinzirl and Fos (30) were among the first investigators to use a 

 very fine sand of standard size. Among the filtering agents which they 

 employed were sands that passed through sieves with 100 and 150 meshes 

 to the inch and mixtures of the former with powdered silica. They also 

 tried varying the depth of the filtering layer from 0.5 to 2.5 cm. They 

 found that, while the mixture of sand and silica was slightly better 

 than the sand alone, sand which had passed through a loo-mesh sieve 

 was very efficient. They also showed that a i-cm. layer of sand was as 

 efficient as a deeper layer. The number of tests recorded was small, and 

 in the minds of some there was still a question as to whether or not lanes 

 or pores that would allow the bacteria to pass through might not form in 

 the sand. This point is discussed further on page 349 in connection with 

 the present studies. The above results of Weinzirl and Fos formed the 

 basis for the recommendation of the sand-filtration method by the Com- 

 mittee on Standard Methods for the Examination of Air (2). 



One of the recent methods of bacterial air analysis which gave sufficient 

 promise of usefulness to be considered by this committee is that of 

 Rettger (22). This method, which is more fully described on page 348, 

 is a modification of Miquel's method of filtering through distilled water 

 (19) and consists of a device for finely dividing the air as it enters the 

 liquid. 



In 1 91 2 a subcommittee (3) gave a report of progress in which they 

 recommended certain modifications in the standard method given in 

 their earlier report. The technique recommended in this later report 



