FUNDAMENTAL UNITS OF MEASURE. 137 



kind of miitter, namely, water, the numerical re[)r('sentation of the ratio 

 being known as " specific gravity." It has taken some years for even 

 scientific men to fully appreciate the objectionable features of this sort 

 of metrology, because it has required soine time to prove beyond doubt 

 that all kinds of copper or silver do not conducli alike, nor do all sam- 

 ples of lampblack radiate alike; and also that the coiulitions nnder 

 which the density of water is constant are difficult of realization. 



Another factor which has been, np to a very recent time, of first 

 importance in the selection of standards is the tendency to seek in 

 nature something of constant dimensions or invariable mass which 

 possesses that general availability essential to adoption as standard. 

 The nomenclature of metrology bears testimony to this. In our own 

 customary system of weights and measures the occurrence of such 

 units as the foot, hand, grain, ell, etc., tells of the frequent recourse to 

 natural units. This is not alone characteristic of earlier and ruder 

 systems, but in modern metrology we have recorded the efforts of sci- 

 entific men to realize this theoretically desirable condition in the selec- 

 tion of the quadrant of the earth, the length of a seconds pendulum, 

 and the wave length of a particular kind of light for linear standards. 

 The oidy natural standard which u}) to this time can be said to have 

 satisfied the requirements is the unit of time, which is the sidereal day. 

 This might itself be considered a derived rather than a fundamental 

 unit, and, indeed, it is difficult to conceive of any time unit other than 

 one based on motion. The motion of the earth is assumed to be a uni- 

 form rotatory motion, and the unit is the duration of a single revolu- 

 tion. Vibratory or periodic motion seems to otter many advantages as 

 a time standard, and various forms have been suggested from time to 

 time. It has been shown that the period of a freely suspended inva- 

 riable pendulum furnishes in practice a more uniform and constant 

 time unit than the best clocks or chronometers. All standards of this 

 type depend on the persistence of gravity, however, and of this we can 

 not be assured. The prime requisites of a standard are constancy and 

 universal availability, and as the present time unit, the sidereal day, 

 possesses these in a high degree it is not likely that it will soon be 

 supplanted. It would be extremely desirable, however, if a unit of 

 time could be devised which would survive such terrestrial or celestial 

 disturbances as would materially alter the revolution of the earth upon 

 its axis. Something of tliis kind is necessary if time observations made 

 during the present cycle are to be available in future ages, and it is 

 ]>ossible that the determination of the relation of the wave length of 

 light to the generally accepted unit of length nniy indirectly furnish a 

 time unit possessing this characteristic in a high degree. 



The greatest advance in the science of metrology in modern times is 

 essentially due to Gauss, and it consists of the so-called "absolute" 

 system of measurement. Quite as much as to the author of this ingen- 

 ious system metrologists are indebted to the celebrated British Asso- 



