Lovering and Bond on Magnetic Observations at Cambridge. 63 
the variation at Boston in 1793 is the mean of 1644 observations. 
The variation at Cambridge in 1782 is deduced from frequent ob- 
servations at different hours on 127 days of that year. Allow- 
ing a reasonable error to beset the value at 1708, we may suppose 
that the number of years from which the rate of annual decrease is 
derived secures it from any large error. The same reliance may 
reasonably be placed in the determination of the rate of increase. 
The fact then is very remarkable that since the passage from a di- 
rect to a retrograde motion the rate of the annual change of varia- 
tion has so materially altered. Assuming that the mean annual rate 
of decrease was 1’.8,* the time of the change interpolated into the 
observations would be 1807; and this in the absence of better au- 
thority may be regarded as its date for Cambridge. 
Another element of the Earth’s Magnetism on which some atten- 
tion has been bestowed at Cambridge is the Dip. All the methods 
of observing the Dip are extremely defective and do not admit of 
so great a degree of accuracy as the Declination instruments. The 
Dipping-needles of Gambey and Troughton and Simms are prefer- 
able to any other direct method of measuring absolute Dip. But 
two of the best Dipping-needles may vary 15’ and more in the de- 
termination of this element for the same time. This difficulty sug- 
gested to Gauss the idea of expressing the Dip as a function of the 
horizontal and vertical components of the magnetic intensity. This 
led to the invention of his Bifilar Magnetometer and the Horizon- 
tal Force Magnetometer of Lloyd. Professor Lloyd has added to 
* In most cases of this kind, the rate varies about the times of maximum 
and minimum as the time, reckoned from these points respectively ; so that 
the value of the element is in proportion to the square of this time. But in 
the present instance, the simple supposition we have made in the text conforms 
best to the observations, 
