THE CONTROL OF THE FACTORY 



547 



as the circumstances will allow, whilst a less degree of exactitude is permis- 

 sible in the first case. Examples of the first case are found in the sampling 

 of the first mill juice, when required to give the executive an idea of the 

 nature of the material being worked up, and in the determination of the 

 purities of material in process made as a guide for regulating the operations 

 of boiling. 



The methods of sampling in use may be defined : — i. Intermittent from 

 a continuously flowing material. 2. Continuously as in i. 3. Intermit- 

 tently from containers, the quantity taken being proportioned to the quantity 

 of material in the container. 



The first method should only be used when general information is required. 

 Th-e second method is accurate provided the sample drawn is proportionate 

 to the rate of flow of the material. The third method is the most accurate. 

 Various methods and devices are described below : — 



If a current of liquid be allowed to impinge on a wire pointing downwards, 

 a very small portion of the liquid will trickle down the wire, and may be col- 

 lected in a container. The quantity drawn depends on the diameter of the 



Fig. 351 



wire. This method is ver}' conveniently used for taking samples of juices 

 from a roller, as when taking first expressed juice, or last mill juice. The wire 

 is supported against the roller and the neck of the container. It is conven- 

 ient to insert a funnel in the opening of the bottle so as to mimimize error 

 from evaporation. Wire sampling may also be readily adapted to juices 

 pumped in pipes, by allo\\ing a jet of juice to impinge on a wire, the excess 

 of juice flo\\ing back to the pump suction. 



Continuous samples may also be taken from pipe lines by employing the 

 arrangement shown in Fig. 351, thus dispensing with the wire. Samples 

 from gutters may be drawn by means of a toy pump. 



A form of automatic sampler described by Maiuice PeUet^ is illustrated 

 diagrammatically in Fig. 352. This is intended to be operated off the miU 

 shaft through a crank, so that at each revolution the bucket dips into the 

 juice gutter, and on its upward motion capsizes its contents into a container. 



A device due to Davoll^ is shown in Fig. 353. A spoon with a channel 

 running through its haft communicates with a hoUow shaft caused to rotate 

 by belt drive from some adjacent machinery. The spoon is covered with 

 gauze so as to keep out fibre. Other gutter samplers are built as undershot 

 wheels, and are caused to rotate by the flow of the juice. They thus auto- 

 matically proportion the sample taken to the rate of flow. A form due to 



