ditions under which the 212° point of the English standard is determined under act of Parliament.
A description of the Kew standards referred to is given in the accompanying table:
| Maker. | Ma ker’s Num ber. |
How Graduated. | Length of 1°. | Small est Gradu ations. |
Length of Tube. | Shape of Bulb. |
| Kew Observatory. . | 585 | 34° to 275° C. | 1·73 mm. | 1 | 618 mm. | Cylindrical. |
| Kew Observatory. . | 578 | 9 to 105 C. | 3·46 " | 0·5 | 455 " | " |
| Kew Observatory. . | 584 | 14 to 220 F. | 1·87 " | 1 | 455 " | " |
The tubes of which the Kew standards are made are about twelve years old, and belong to the series purchased by the Royal Society and deposited at Kew to be used as standards.
The essential parts of the water comparator in use at Yale for comparing thermometers is shown in outline in Fig. 1, where a a' a" is a bright-tinned iron cylindrical tank 15 X 20 inches, having an aperture for the stopcock f, a lid a a" with various apertures for the insertion of long thermometers, and having a plate-glass window 4 x 14 inches set in the side. Within this outer tank a smaller copper tank 11 x 15 inches is symmetrically placed, and rests upon wooden bars which are supported by the bottom of the outer tank. A window, placed in the same relative position as the outer window, allows the thermometers which are attached by springs at their upper ends to the adjustable brass disk d', to be read. The brass axis c c' turns in a bearing c', and has attached to it two disks, d, d'. A small cathetometer with its telescope is placed before the, windows, and the number of the thermometer under observation is shown by means of the graduated dial at d. Water of a given temperature is admitted through the tubes at e, and after the temperature has been brought to the degree required, it is thoroughly agitated by moving vertically the ring plunger shown at p p p' p'. The disk d', which is perforated, will accommodate sixty-four thermometers. The agitation of the water having subsided, the thermometers to be compared are read as rapidly as possible, first from left to right and then from right to left. Two standards are read at the beginning and end of the series. It is obvious that, if they are read at approximately equal intervals, the mean of the two readings will be free from the error of radiation caused by the slow cooling or warming of the water. The greater part of the work of the observatory upon standards is done with this comparator. For clinical thermometers a smaller apparatus, constructed on the same general principles, is used; but, as in this case a much less degree of precision is desired than in the investigations of standards, the work may be simplified. It is not necessary, for instance, to resort to the somewhat tedious read-
