In 1973 the first volume of An Historical Geography of Europe was published. It presented a survey of the geography of the continent at five widely separated periods of time, from the fifth century B.C. to the fourteenth century A.D. It stressed what has come to be called the ‘horizontal’ approach, and gave little space to the evolutionary processes which occurred between the periods chosen. Such a method is open to criticism. The choice of the periods for intensive study is not automatic, and several periods of great interest and historical significance received no consideration. The choice must be arbitrary, unless regularly recurring periods are chosen – every five hundred years, for example. Such a method would have been very difficult to use, because some periods, the fifth and sixth centuries A.D., for example, would be quite impossible to document on the scale contemplated.

The author tried to use what might be called ‘peak’ periods in European history, the culminations of long historical processes. At one time he considered using those examined by the art critic Clive Bell in his study of Civilisation. The fifth century B.C. was an inevitable choice; perhaps also the Carolingian period in the late eighth and early ninth centuries, but the age of the Flavian emperors hardly marks the climax of the Roman empire. Most would regard the thirteenth rather than the fourteenth century as the culmination of medieval civilisation, and the period around 1100 was chosen because a survey seemed to be needed between the ninth century and the fourteenth.