According to the vegetation investigation and pollen analysis of surface samples sampled along a precipitation gradient of the Northeast China Transect (NECT), several pollen taxa, including Pinus, Betula, Quercus, Tilia, Acer, Ulmus, Artemisia, Chenopodiaceae, Gramineae and Cyperaceae, were chosen to make the regression and correlation analyses. The results indicated that there exists a close relationship between vegetation and pollen taxa in surface samples. The regression parameters for ten taxa in the forests in the eastern part of NECT were different from those in the steppes in the western part. Pinus, Betula, Artemisia and Chenopodiaceae, which have large slope and y-intercept terms, were over-representative taxa. Acer, Gramineae and Cyperaceae, which have small slope and y-intercept terms, were under-representative taxa. Quercus, Tilia and Ulmus whose slope terms have negative correlation with y-intercept terms were equi-representative taxa. The pollen taxa with large slope or large y-intercept terms have small variability coefficients, implying that the slope and y-intercept terms for these pollen taxa are of high accuracy in the estimation of plant abundance from pollen frequencies.
Keywords: pollen taxa in surface soil, vegetation, regression parameters (slope term and y-intercept term), representation, Northeast China Transect (NECT).The fossil pollen records are important for reconstructing palaeobiome in global change studies, but it is very complicated to transform fossil pollen data to the accurate estimation of past vegetation [1] . The quantitative relationships between the vegetation and pollen taxa, including Pinus, Betula, Quercus, Tilia, Acer, Ulmus, Artemisia, Chenopodiaceae, Gramineae and Cyperaceae, are essential in reconstructing the past vegetation history in the Holarctic region [2,3] . The former studies mainly focused on the arboreal pollen; little has been done on the herbaceous pollen and the comparison of quantitative relationships between pollen and vegetation among different vegetation types [4,5] . Based on the vegetation investigating and surface sampling along the Northeast China Transect (NECT) [6] , using correlation and regression analyses as tools, we studied the quantitative characteristics of several common pollen taxa and compared the regression parameters of a pollen taxon sampled from different types of vegetation. The application of the results in reconstructing the vegetation history is discussed.
Methods( ) Vegetation investigating and surface sampling.The NECT, extending from the Changbai Mountains in Jilin Province to Erenhot in the Inner Mongolia Autonomous Region, is situated between longitudes of 112°E and 130°30′E and latitudes of 42°N and 46°N. The transect was set along the central line of 43°30′N. It is 1 600 km long from the east to the west and 300 km broad from the south to the north, and has a strong east-west gradient of precipitation. The global position system (GPS) was used for location in the field survey. 30 stands were chosen for vege...