420 EVENING DISCOURSES. 



Admiralty compass experts in all countries. The lecturer considered that he 

 was doing general good, and was not doing himself much harm in disclosing the 

 knowledge he had spent so much time and money in obtaining. 



The greatest of these errors was what he called the quadrantal error ; its 

 elimination has had more to do with drastic changes of design in all forms 

 of gyro-compass than anything else. When a ship's course is neither N. nor S. 

 nor W. nor E., but between these directions, and particularly N.W. or M.E. 

 or S.W. or S.E., and if she rolls steadily, the compass gradually gets away 

 from the north, setting itself eonietimes with an error of 20°. Mr. Brown's 

 popular explanation of this phenomenon and its remedy are too long for an 

 abstract. They may be briefly put as follows : It has long been known (see 

 Evans and Smith, 1881, Phil. Trans.) that to avoid this error there must be 

 dynamic balance {that is the moment of inertia A of the gyro about OA, the 

 axis of the wheel, must be equal to the moment of inertia B about OB, a horizontal 

 axis at right angles to the first). ^ Auschiitz certainly knew this in 1911, but he 

 tried in vain to get dynamic balance, and his use of many gyroscopes in one 

 compass was an attempt to get round the difficulty, but it resulted in great com- 

 plexity. Mr. Brown saw that it was impossible to get dynamic balance, because 

 although B was fixed and definite, A depended upon the rate of rolling of the 

 ship. If there were absolutely no friction at the pivots of the wheel the inertia 

 of the wheel would form no part of A and balance would be possible, but there 

 is always some such friction, and the inertia of the wheel is more or less added 

 to A, so that A is variable and its balance is impossible. But a careful study 

 of terms usually, neglected in the mathematical theory showed that whatever A 

 and B might be, there would be no quadrantal error if the vibratory angular 

 motion about OA could be made to differ in phase from the motion about OB 

 by a quarter period. Before he had a mathematical reason for this he had 

 constructed a form of gyro-compass quite different from what he had hitherto 

 been trying, and by intuition and good fortune he found that it had no 

 quadrantal error. Later, he found that it fulfilled this mathematical condition. 

 But the intuition and success came before the mathematics. It is always wise, 

 however, to get as good dynamic balance as possible by a symmetrical disposition 

 of masses, and also to minimise both the angular motions. This discovery of 

 the cause of tihe quadrantal error and its correction are here made public for the 

 first time. 



Besides well-known simple apparatus, Mr. Brown exhibited his master c&mpass 

 which had been turned quite away from the north before the lecture. There 

 were two very visible repeaters, and there might have been many more, which 

 faithfully copied the master compass. One of these repeaters drew a well- 

 damped curve showing how the master compass was gradually getting into its 

 proper north position where it would remain so long as the wheel rotated. 

 One of the repeaters, called the steering repeater, had a compass card in which, 

 near the lubber line, the divisions of the card were four times as large as usual ; 

 in fact, one degree was really, represented by four degrees, giving four times as 

 much accuracy in steering, besides giving less strain on the eyes. Its construction 

 and the construction of the other exhibits were easy to understand by mere 

 inspection. The lecture was illustrated by many lantern slides. As the readers 

 of this abstract cannot see the master compass (Mr. Brown exhibited one detached 

 from its binnacle) here is a brief description. Instead of letting gravity, act as 

 with a pendulous weight, Mr. Brown has two bottles connected by a horizontal 

 tube and containing oil. The wheel in its case is really a blower, and air comes 

 under pressure to a valve to establish a pressure difference between the bottles, 

 which is proportional to the tilt, so that the difference of weights of oil in the 

 bottles is proportional to the tilt. The action is just the same as that of a 

 pendulous weight. Anotlier, smaller, pair of bottles is acted upon in just the 

 opposite sense, so as to oppose the first pair, but the tube connecting these 

 smaller bottles is greatly constricted; it results that the oscillations are damped, 



^ Mr. Brown illustrated dynamic balance by suspending a child's hoop by a 

 string and swinging it to the north and south. The hoop sets itself with its 

 plane east and west. Then he fixed a similar hoop at right angles to the firet 

 and swung the combination to show that it did not tend to set itself in any 

 direction. It is now in d,vnamic balance. 



