If the observer desires to secure an exact adjustment of the zero of the scale, or in any case if the latter deviates more than one-half of a degree, the zero lines are made to coincide by moving the milled head and securing a neutral field at this point by means of the small key which comes with the instrument, and which fits into a nipple on the left hand side of F, the fixed quartz wedge of the compensating system. This nipple must not be confounded with a similar nipple on the right hand side of the analyzing prism, H, which it fits as well, but which must never be touched, as the adjustment of the instrument would be seriously disturbed by moving it. With the key on the proper nipple it is turned one way or the other until the field is neutral. Unless the deviation of the zero be greater than 0.5 deg., it will not be necessary to use the key, but only to note the amount of the deviation, and for this purpose the observer must not be content with a single setting, but must perform the operation five or six times, and take the mean of these different readings. If one or more of the readings show a deviation of more than 0.3 deg. from the general average, they should be rejected as incorrect. Between each observation the eye should be allowed 10 to 20 seconds of rest. The "setting" of the zero having been performed as above, the determination of the accurate adjustment of the instrument by means of the "control" quartz plates is proceeded with. Three such plates will be furnished with each polariscope, which have "sugar values" respectively approximating 80 deg., 90 deg., and 100 deg.. These values may vary with the temperature, and tables are furnished with them which give their exact value at different temperatures, from 10 deg. to 35 deg. C. One of these plates is placed in the instrument, and the field observed; it will be seen that the uniform appearance of the field is changed. The milled head is turned to the right until the exact point of neutrality is re-established, just as described above in setting the zero. The scale is read, the observation repeated, the reading taken again, and so on until five or six readings have been made. The average is taken, readings being rejected which show a divergence of more than 0.3, and the result corrected for the deviation of the zero point, if any was found, the deviation being added if it was to the left, and subtracted if to the right. If the adjustment of the instrument be correct, the result should be the value of the control plate used, as ascertained from the table, for the temperature of 20 deg.. Each of the three plates is read in the instrument in this way. A variation of 0.3 from the established values may be allowed for errors of observation, temperature, etc., but in the hands of a careful observer a deviation greater than this with one of the three plates, after a careful setting of the zero, shows that the instrument is not accurately adjusted.