OCEAN TEMPERATUEES. 161 



mometer has accidentally beeu broken, the observations have been coutiuned by means of a light- 

 house thermometer, or of one purchased at a neighboring town, pending the receipt of a new 

 instrument, but such instances have been of rare occurrence and not likely to make any appre- 

 ciable difference in the character of the records for the short periods involved. 



While the light keepers have not been specially trained in the methods of taking temperature 

 observations, their monthly returns testify to their high gi'ade of intelligence and to their great 

 zeal in fulfilling these additional duties without extra compensation. We are led to believe that 

 their records contain comparatively few serious errors for which they are directly responsible, 

 excei^tiug in those cases which will be explained hereafter; and that, in the main, their observa- 

 tions have been conscientiously made and are deserving of consideration. It is also worthy of 

 comment that so extensive an undertaking as this should have been carried on successfully at so 

 little expense and with so little friction. 



As above explained, detailed temperature observations rather than general results are essen- 

 tial for comparison in studying the migrations of the tishes. For that reason the reductions plotted 

 on the accompanying charts have been made for comparatively short intervals, the entire year 

 being divided into periods often days, each of which is equivalent to about one-third of a calen 

 dar month. As two observations are made daily, the mean of each ten days is derived from twenty 

 observations, and small errors are thereby practically eliminated. Each station is represented by 

 a chart on which the ten-day means of surface-water temperatures are given for each of the five 

 years from ISSl to 1885, inclusive, and the air temperatures from 1881 to 1883, inclusive. The 

 method of representing the temperatures is by curves connecting the ten-day periods, as explained 

 on the charts. At the stations north of Cape Hatteras there were frequent indications of careless 

 observation during exceedingly cold weather in the months of January and February, the ther- 

 mometer, at times, not being read quickly enough after it had been withdrawn from the water. 

 As such readings would manifestly afford a lower mean temperature than the actual, it has been 

 deemed best to omit the records for those two months at the northern .stations. At some of the 

 shore stations the observations also show the effects of local influences which render them inap;- 

 plicable to the open waters of the coast; but special explanation of those stations is made furtlier on .. 



In addition to the charts of tendaj- means, there are also seven isothermal charts on wliichi 

 the temperature observations at all the stations are combined, in order to afford more convenient, 

 means of comparison. Five of these charts represent the separate years from 1881 to 1885, inclu- 

 sive, a sixth, the means of the same five years, and the final one, the relations of the air isotherms. 

 to the surface isotherms. The isotherms are plotted for every 5° of temperature, Fahrenheit, from, 

 40° to 80°, inclusive. The isotherm of 35° F. occurs only at the northern stations, and there mainly- 

 iu the months of January and February, the records for which have not been used. The writer 

 has refrained from drawing any conclusions from the temperature results presented in this leport, 

 and his remarks on the following pages are mostly confined to describing the stations and indicat- 

 ing some of the main features with respect to the temperature curves and the isotherms. The 

 Fahrenheit scale of temperatures has been exclusively used both in making the observations and in 

 the construction of the charts. 



The work of reducing the many observations to ten-day means and of making the original 

 l)lottings of the same has been done by Miss INI. J. Eathbun, while the writer is respi)usible for 

 the computations for, and the plottings of, the isotherms. The charts were prepared for engrav- 

 ing by Mr. C. E. Gorham. 

 SEC HI 11 



