10 Fishery Board for Scotland. 



particular line in space (say longitude or latitude) for all the twelve 

 months of the year. Working with such data as are available in the 

 present case, we usually find that, in making such an isopleth diagram 

 from the twelve sets of data furnished by our monthly charts, the 

 points are found to be on the whole very concordant, and the drawing 

 of our isopleth lines is easy enough ; but at the same time a little 

 further smoothing is in most cases necessary. And after this is done, 

 we may once more go back to our isotherm charts, and make such 

 slight alterations upon them as our isopleth diagrams in turn suggest. 



Before we leave this part of our subject, it is worth while to call 

 attention to a certain difficulty which constantly besets us in the 

 making of charts for the narrow seas, such as the Chamiel or the 

 North Sea, from such data as we have to deal with. These data are 

 of two kinds — data from shore stations and data from observations 

 at sea. On the whole the shore data, from observations made at 

 lighthouses, harbour stations, or inshore light-vessels, are the best we 

 have, for they have been made daily, or even oftener, over long 

 periods of years ; from these inshore observations alone we can draw 

 very fair temperature charts for our whole British area, and even, 

 with the help of similar continental observations, we may, by inter- 

 polation and extrapolation, continue our isotherms across the sea. 

 But while such charts would be found to correspond very fairly well, 

 in a general way, with those which we draw from our actual 

 observations at sea, there is always a very great ditticulty in bringing 

 the two sets of observations together in their minuter details. The 

 inshore temperatures have features of their own ; neighbouring 

 shore stations may differ much from one another in their proximity 

 to large rivers or in other ways ; and, as we shall see presently, 

 proximity to land has a very marked effect upon the sea temperatures, 

 and this effect is apt to be very different at one season from another. 

 Now, as what we are anxious to understand and to represent upon 

 our charts is not the shore temperature but the general surface 

 temperature of the ojjen sea, it follows that a great part of our 

 available material, and a part which in other respects is the most 

 trustworthy that we possess, is ill adapted for our particular 

 purpose. The dif^reuces are not of great magnitude, and we can 

 overcome the difficulty, more or less completely, in various ways. 

 But the fact remains that it is not possible, without a certain amount 

 of " smoothing " or averaging, to draw systems of isotherms which 

 shall clearly illustrate our main temperature phenomena, and which 

 shall keep with equal closeness to the data furnished by our inshore 

 and our offshore observations. 



It may seem at first sight that these successive manipulations of 

 our original data must lead us further and further away from such 

 accuracy as the actual observations contained. But this is by no 

 means so. We are not dealing, as we sometimes profess to do in a 

 laboratory experiment, with a set of data each one of which is as 

 " a nail in a sure place." On the contrary, we are dealing with a 

 vast mass of figures, each one of which is only an approximation, and 

 not even a very close approximation, to a true statement of the mean 

 temperature at a certain place and time ; and it is in the inter- 

 connected web of all these approximations that the truth which we 

 are seeking lies. The more, then, that we can associate and correlate 



