CHAP. 34] THE PLEISTOCENE RECORD 897 



The equatorward limit of perennially frozen ground, which is believed to lie 

 close to the mean annual isotherm of — 2°C, probably affords a better (although 

 still imperfect) basis for inference as to secular climate. On the assumption that 

 the limit at the time of the last major glacial maximum has been correctly 

 traced across Europe from geologic data (Poser, 1948), mean annual tempera- 

 tures 8°C to 10°C below existing values can be inferred. The chief uncertainty 

 here lies in the reliability of identification, from relict features, of former 

 frozen ground. 



Fluctuations of lakes in basins without outlet are dependent not only on 

 temperature but on precipitation as well. A crude value (probably maximum) 

 for mean annual temperature during the last high-lake episode in the Basin- 

 and-Range region of western United States was found by Meinzer (1922) to be 

 8°C less than today's values. 



As a group, terrestrial plants are thought to be more sensitive to temperature 

 changes than are terrestrial mammals. Consequently special attention has been 

 given to Pleistocene fluctuations of the ranges of fossil plants. In Europe, where 

 most of the studies on record have been made, we find derived mean annual 

 temperatures 4° to 6°C lower than today's at glacial times, and at one locality 

 2° to 3°C higher than today's at an intergiacial time. We find also comparable 

 values exj)ressed as July means. 



Three nearly continuous paleobotanical sequences give a picture of tempera- 

 ture fluctuation during earlier parts of the Pleistocene. One is the lacustrine 

 sequence at LefFe, northern Italy (Lona, 1950; Venzo, 1955; Lona and Follieri, 

 1957), reproduced in its essentials in Fig. 1. The altitude of the locality is about 

 400 m. Temperature fluctuation equivalent to a change of altitude of more than 

 2000 m within the full glacial-interglacial range is implied. The apparent mean 

 annual temperature at the coldest times shown by the curves in Fig. 1 was 

 about 9°C lower than that of today (Lona, in litt.). The second is the sequence 

 (Szafer, 1954) from Carpathian Poland, with a rich content of megafossils, 

 extending from well back in the Phocene to a point that is believed to lie near 

 the middle of the Pleistocene. From the Pleistocene part of the sequence, 

 deduced mean annual temj)eratures range through 13°C ( + 8°C to — 5°C), 

 the lowest value being based on Arctic-Alpine tundra recording an extensive 

 glaciation at the top of the measured sequence. The highest value is close to that 

 of today in the same district. Because of the continental position of the district 

 concerned, the range of temperature fluctuation may well be substantially 

 greater than that of the secular mean. It does not yet seem possible to use the 

 Polish data to extend the curves shown in Fig. 1 further up into the Pleistocene, 

 but the data deserve close study. 



Some of these pollen sequences and other sequences having climatic implica- 

 tions for the last several tens of thousands of years were compared by Flint and 

 Brandtner (1961). The agreement among them is fairly good, suggesting general 

 synchrony of temperature fluctuation during the time represented. 



The third is the sequence from the Netherlands, developed by Zagwijn (1957), 

 mainly from pollen data. It agrees fairly well with that from Lefife as to 



