INVESTIGATION OF TEMPERATUEE CHANGES. 229 



during which accurate observations have been taken is not sufficient ; and 

 most of the longer series of records of temperature come from observatories 

 situated either within or very near to cities where the conditions have not 

 remained the same, but have been rapidly changing, and in such a way, 

 we have good reason to believe, as to produce a decided eft'ect on the tem- 

 perature. Tlie most satisfactory result is that from the cellars of the Obser- 

 vatory at Paris ; but even here we are not certain that all sources of error 

 have been avoided. If they have been, Ave may infer that the change of 

 climate at Paris, so far as temperature is concerned, has not been sufficient 

 in fifty years to sensiblj' aflect the thermometric mean. 



If, as appears to be the case, instrumental observations have been carried 

 on for too short a time, and too imperfectly made, to justify us in expecting 

 that their records should furnish positive information in regard to any sus- 

 pected change in the temperature of the globe, the question arises, whether 

 trustworthy conclusions cannot be drawn from data obtained in some way 

 not connected with the use of instruments, but leading to results of a similar 

 character to those which might be furnished by a lai'ge number of instru- 

 mental observations made in many localities and with a sufficient degree of 

 precision. For instance, historical records of important events connected 

 with climate, changes in the distribution of cultivated plants, changes in the 

 character of forests, migrations of nations, dying-out of the inhabitants of 

 any particular region, — such occurrences as these might be shown to be of 

 such a nature as to be accounted for only by supposing or admitting that 

 the climate of the country from which the evidence came had undergone a 

 change during the corresponding period. It Avill readily be seen that there 

 are many occurrences of sufficient importance to the welfare of the people 

 to be handed down in their traditions or historic records, and which might 

 be legitimately used in the manner thus indicated. An investigation of this 

 kind, however, opens too wide a range of discussion to permit of any thor- 

 ough working up of the subject. The most that can here be done is to give 

 the general results at which the present writer has arrived, after examining 

 with considerable care all accessible authorities.* 



By far the most important and elaborate publication relating to the ques- 

 tion whether there is evidence, not instrumental, of a change of temperature 

 during the historic period in any part of the globe is that of Arago, to which 



* The author hopes to be able, at some future time, to set forth more fully than can here be done the various 

 facts collected by him in reference to evidence of changes of climate in historic times. 



