This paper is based on identification of the pattern of the upper level of the world city network of knowledge as published in a series of earlier papers. It is our aim to update the findings and relate to the general world city discussion. The structure of the world cities of knowledge network has changed over the past decade in favour of south-east Asian and south European cities and in disfavour of the traditional centres of North America and north-western Europe. The analysis is based on bibliometric data on the world’s 100 largest cities measured in terms of research output. The level of co-authorship between researchers in different cities is an indicator of links and respect, and the number of citations of papers produced by researchers located in each city is an indicator of respect. Finally, one research discipline is selected for an experiment in forecasting future hot spots of research.

Chinese scientific output has increased dramatically in recent years, but its internal spatial structure has received scant attention. Estimated gravity models of intercity scientific coauthorships show that there are two types of spatial political bias in China, apart from the expected mass and distance effects. Intercity coauthorships involving Beijing are more common than Beijing's output volume and location would imply, and this Beijing bias is increasing over time. The second type o f spatial political bias is greater intraprovincial collaboration than is accounted for by size and distance. The geography o f Chinese science is thus not only monocentric as regards overall scientific output, but also exhibits unusually hierarchical collaboration patterns. Unlike in Europe and North America, national and regional capitals are becoming ever more important as scientific coordination centers.

The importance of the knowledge-base in regional and urban competition is generally recognised, although causal relations between urban and regional economic growth and knowledge level are far from clear. This paper presents the rst analysis of the strength, interrelations and nodality of the global research centres. The data are records in the Science Citation Index 1997-99 of papers produced by authors from the 40 largest 'greater' urban regions of the world as measured by research output. The network of research co-operation depends on nationality, distance and other factors. The top-level nodes in the co-operation network of the world are London, Genève-Lausanne and the San Francisco Bay Area. In absolute number of co-authored papers, Los Angeles, Boston and New York constitute a second level and, when observed links are related to expected links, the second level combines Amsterdam-Hague-Rotterdam-Utrecht, Paris, Basel-Mulhouse-Freiburg and Copenhagen-Lund. As expected, the networks of citation are, by contrast, very independent of distance, but not of nationality. The primary categories of research centres for the total number of citings presented are San Diego, Seattle, Boston, New York and the San Francisco Bay Area. When we turn to the international data-set, it is Mannheim-Heidelberg, Genève-Lausanne, Basel-Mulhouse-Freiburg and Cambridge which are in the lead.The declining importance of physical distance has given nourishment to a widespread belief that the concepts of place and territory are without meaning in a globalising society and economy. When it comes to research and knowledge, it is often stated that research can be carried out without any constraints of place or territory and that knowledge can be moved freely. We have the opposite opinion-namely, that research and knowledge come out of local culture and tradition and have a strong territorial af nity. In line with this, we nd the territorial base to be very important for the production and distribution of knowledge. In this paper, we analyse the global top level of research centres as measured by total research output. We present the major research metropoles of the world in terms of research strength, interrelations and nodality within their part of global research networks.The present paper is a development of Matthiessen and Schwarz (1999), widening the region of analysis from Europe to the Christian Wichmann Matthiessen is in the Institute

No abstract