TKAXSACTIONS OF SECTION A. 449 



3. The lamps and motor should have a high resistance, the standard lamp not 

 higher than 475 to 500 ohms. 



4. The resistance of the conductor should be as low as practicable, and should 

 be graded in size for the current required. 



5. The electromotive force of the generator should never rise to over 170 volts. 



6. The lamps should be simple of construction, durable and cheap, and all 

 alike. 



7. The generator and the engine should be of the most substantial construction. 



8. The points of derivation should all have, and all be maintained at, the same 

 difference of potential as nearly as possible. 



9. The energy necessarily lost in armature and conductors of any fixed resist- 

 ance for lamps of the same absolute economy varies inversely as the resistance of 

 the lamps, or as the squares of the currents used. 



10. Since in a single generator the total power used is not proportional to the 

 number of lamps in circuit, a machine should be run nearly to its maximum 

 capacity. 



11. There should not be a single generator or a single circuit which maybe 

 destroyed ; but there should be such an arrangement that the external demands, 

 both with regard to resistance and current required, may be met by suitable 

 changes in the resistance and supply of the generative system. 



12. The relation of internal and external resistance should be in a system the 

 same as exists in a single generator when worked to nearly its full capacity. 



13. With such proper relation of internal and external resistance the electro- 

 motive force will vary but slightly, and the power used will be very nearly pro- 

 portional to the number of lamps or the current developed. 



14. Insulation of conductors and of the armatures should be good, and provision 

 made to prevent fire arising from an abnormal increase of current. 



15. Means should exist of accurately measuring the current used. 



IG. A large system should be as economical as a single generator worked 

 nearly to its full capacity, and should be capable of regulation with the same ease. 



Much has been said about the subdivision of the electric light and the subdivision 

 of the current ; and such subdivision of electricity has been the great bugbear in 

 the distribution of power and light by its means. This term, while conveying 

 an idea, is in reality essentially wrong. It implies the existence of a current, 

 the expenditure of energy, before the required subdivision of such current is made. 

 A current is a function, not only of potential, but of potential and resistance. 

 We desire to establish a certain potential, or difference of potential, and to maintain 

 this so that when a circuit is established, a channel opened, a current may be 

 established in that circuit. We do not divide a current, do not divide a potential, 

 but having established a difference of potential in two conductors, we open one or 

 more paths, and a current flows over these paths dependent on their resistance, 

 and independent of each other, and such additional current having been formed, 

 the current over the main conductor is increased by just such increment, which 

 increment did not exist until the new path was made. There has been no sub- 

 division of a current, there has been a creation. Such new current havino- been 

 made, of course more power must be supplied to maintain the existing poten'tial. 



7. 0)1 the Comimrison of the Mercur// with the Hydrogen Thermometer. 

 By Professor J. M. Cbafts. 



A thermometer filled with air or some other gas is the standard instrument for 

 measuring temperature, and a correction is necessary to bring the results of a 

 mercury thermometer into accord with this standard. The only table of corrections 

 for temperatures above 100 degrees Centigrade is that published by Eegnault 

 thirty years ago. It has been found that this does not apply to the instruments in 

 use to-day, and a new series of determinations has been made of a number of 

 French and of one German thermometer, giving results so concordant that a new 

 table for general use can be founded upon them. 



The deviation of the mercury from the gas thermometer is much smaller than 

 1882. G G 



