fourth decimal place. Ordinary distilled water, or even boiled rain water, is sufficiently good for the purpose.

A specific gravity bottle is the only trustworthy way of getting moderately accurate specific gravities, and whether home-made or purchased, it is easily tested.

Hydrometers marked for s. g., and for degrees of any arbitary scale, as Baumé, Cartier, etc., are useful for many purposes and very convenient in application, but their results are at best only rough approximations. For frequent extemporaneous use in unskilled hands they furnish very useful information in a very short time, and with very little trouble, if so constructed and managed as to give their best indications. But all the old arbitrary scales should be avoided, and those marked for specific gravity only should be used.

THE URINOMETER.

This instrument is simply a small hydrometer with a range or scale limited to the specific gravity of urine. As commonly sold it is so faulty and so unmanageable as to be almost useless, while the need of a moderately trustworthy instrument is daily more and more felt, as accurate observation is so rapidly improving the art of medicine. Next to the clinical thermometer there is perhaps no instrument more useful to the observant physician, nor more frequently required in daily practice, than the urinometer. In the importance of its indications, also, it is often second only to the thermometer. The specific gravity bottle is the only accurate urinometer, and the foregoing note on the specific gravity of liquids was written mainly as introductory to the use of specific gravity in the examination of urine. But the specific gravity bottle, and the necessary scale and weights cannot be carried to the bedside, while a properly constructed hydrometer can be, and will serve well to indicate when the greater accuracy of the specific gravity bottle is needed. With a good hydrometer and a little skill and practice in the use of it the bottle will not be needed once a week in an active practice, and yet all the needed information will be had with very little loss of time and without the carrying of vials of urine oftener than is needed for microscopic examinations.

There are some good urinometers sold, but the number is small, and often the best are so large that they are not adapted to the pocket, and often it is impossible, when they are most needed, to get enough urine to fill the jar, while their required temperature of 60° F. is difficult and troublesome to attain. But by far the largest number sold are grossly inaccurate, and almost impossible to manage within the narrow limits of their utility.

Their defects are numerous and very general, but the defect which chiefly invalidates them is the cylindrical shape of the air chamber as seen at Fig. 4 of the following illustrations. This shape in a cylindrical jar but little larger than the hydrometer gives a line of contact and cohesion between the instrument and the inside of the jar which is fatal to a free movement up and down in the liquid, and therefore a free and correct floating is nearly impossible. It is quite impracticable to keep the instrument in the centre of the liquid even for a moment or two until read, and the moment it touches the jar the reading is so liable to serious error that the instrument may become not simply useless, but hurtful through giving information that misleads. Beside this, the stem is usually too large to give a distinct reading scale. A proper shape for the air chamber is that of a double cone, base to base, as seen in Fig. 1. In this there is but a single point of contact with the side of the jar, and thus the friction and cohesion is limited to the minimum degree. The stem should be as small as possible to contain the paper scale, because upon this depends the width of the subdivisions of the scale and the distinctness of the reading. The scale should extend from 1.000 to 1.060, and should be so divided that each figured line should represent 010, and each subdivision 001, or a unit in the third decimal place. The scale is sometimes marked without the decimal point, as 1000,-1010, etc., and sometimes the unit figure, decimal point and first 0 are omitted, and the scale simply reads 0,-10,— 20,-30, etc., the subdivisions being always the same. These latter forms of scale are good enough and are easily understood, but they are not the best form, and are not accurate in expression, for it is a specific gravity instrument, and water being unity, the other divisions are decimal fractions of unity, and should be so expressed. The scale should indicate the temperature to which it is adjusted, and this temperature should not be 60° F. or 62° F., as is common, but should be 25° C. 77° F. This latter is a much more convenient temperature because it is nearer to that of the modern sick room, and because if below that it can always be attained by the warmth of

=

the hand on the outside of the jar in a minute or two. And if the temperature of the room in which the urine has been standing is within two or three degrees of that either way it is sufficiently near. The urinometer being adjusted to read 1,000 in water at 25° C.= 77° F., the reading will be one unit of the third decimal place, or one subdivision of the scale, too high for each 3° C.=5.4° F. above that figure. Therefore if the urine be measured just after it is voided, say at 37° C. 98.6° F., it will indicate just '004, or four subdivisions of the scale above the true s. g. This of course is the highest temperature at which urine could be measured, and if such urine should indicate a s. g. of 1.026, and the '004 be subtracted from it for excess of temperature above that of the scale, the true indication would be 1.022. If a specimen of urine was tried at 3° C.=5.4°F. above the scale temperature, that is, at 28° C.=82.4° F., then it would be 001, or one subdivision above the true indication, and so on. It will thus be seen that four subdivisions of the scale is the greatest possible error from excess of temperature that this scale admits, and that being within the limit of variation of normal urine (about 1.018 to 1.028) is not a very serious matter in a large proportion of the applications of the instrument. Hence it is one of the advantages of this scale temperature of 25° C.=77° F. as a standard, that in a very large number of the instances of the very useful though negative applications of the urinometer, a thermometer is not essential, and is only needed when some abnormality shows the necessity for closer observations.

The next modification that the writer has to suggest is perhaps a still more important one, because it renders the use of the instrument easier, with far less liability to error. This improvement is in the jar, and consists in indenting the sides of the jar, as represented in Fig. 1, and in cross section by Fig. 2. The jar is cylindrical, as in common urinometers, but has a foot for steadier standing. The glass of the cylinder is softened in the lamp, and while soft is indented so as to make a V-shaped projection inside the jar. These longitudinal grooves or indentations are repeated, so that there are four in all, and the effect is that the hydrometer cannot come to the sides of the jar at all, but must either float free in the centre or touch one, or at most two of the sharp edges of these indentations, as shown in the section, Fig. 2, and where it does touch, it can only be by a single point of contact between two convex surfaces, and thus the friction and cohesion will be reduced to a minimum. The jar, as seen at Fig. 1, is made by the lamp, and the flutes or grooves

only occupy the middle half of the cylinder, but if made in a mould the grooves would extend from the lip to the foot, and this would be better, as then the jar would require still less urine to fill it. This is an important consideration, because it sometimes happens that the urine is very scanty when it is most important to know its density or s. g. All the apparatus is shown of full size in the illustrations, and the jar holds about 45 c. c., or 11⁄2 fluidounces, and when the urine is above the normal density a fluidounce or 30 c. c. will fill the jar sufficiently to float the hydrometer. If the indentations extended from top to bottom, as they should, somewhat less urine could be used.

The reading of the indications on the stem is a matter of importance, and requires the personal observation of each individual, and each will fall into a habit of reading, that will be sufficient, if adhered to, as long as the instrument lasts, but a good habit of reading is as easily fallen into as a bad one. Many persons read the indications from below the surface of the liquid, but this is not a good habit for two principal reasons: First, all hydrometers are constructed to be read from above the surface, because they are often used with colored and opaque liquids, which prevent reading from below; and secondly, because the distortion by the liquid is not inconsiderable, and varies with the density of the liquid. Therefore the indications should be read from above the liquid, or at least from above as well as below. The first thing essential to a moderately close reading is to have the stem entirely clean, for if greasy or soiled by the sebaceous matters of the fingers from handling, the liquid will not touch the stem in the proper manner for reading. When clean, the liquid piles up against the stem, covering from two to three divisions of the scale, and the rise of the liquid from the general level of the surface up against the stem gets thinner as it rises higher, so that the absolute point or feather-edge to which it rises cannot be seen, but there is generally a spot or section of a line of reflected light from the concave surface of the liquid when the stratum gets very thin upon the stem, that is useful as the line of proper reading, and the subdivision of the scale, which is nearest to this point when the instrument has come to rest, is the proper one to read. But very often this spot of light cannot be seen, or is seen as a broad and illdefined reflection; then let the instrument be pushed down a little and set to vibrating up and down, and while it is thus moving let the line of liquid be observed against the moving stem. The horizontal line of the liquid does not move, but the stem does, and by