POLLEN AND SPORES — LEOPOLD AND SCOTT 315 



genera because, though one might be sure to what major taxonomic 

 group the plant belongs, it is usually considered inappropriate to 

 apply the names of living genera to such old material. 



Identification of fossil forms in terms of modern species and genera 

 requires careful comparison with a large-as-possible collection of 

 modern pollen, prepared by acetylation from authentic herbarium 

 collections (Traverse, 1955). 



With proper preparation a fossil pollen sample may contain up to 

 1,000 grains per slide. If the assemblage represented is rich in types, 

 it may consist of 60 to 100 forms, though usually it contains less. An 

 estimate by eye of the relative proportions of all these types is usually 

 not accurate, especially for the rarer forms, and instead it has become 

 accepted practice to count 200 to 1,000 grains in order to compute the 

 percent composition of a sample. Because the slide assemblage is 

 mixed, systematic traverses of the preparation by means of a mechani- 

 cal stage permit the observer to encounter a random sample. By 

 performing counts in a consistent manner for each of several samples 

 in a sedimentary sequence, and by converting the tallies to percent 

 composition of the observed sample, quantitative data can be obtained. 

 By plotting the data in graph form with the values for each sample 

 arranged in a vertical series according to its placement in a section, 

 the relative numerical importance of each pollen type at different levels 

 in the section can easily be seen. Such a plot is termed a pollen dia- 

 gram, an example of which is included as figure 6. 



When the sample count includes 1,000 grains, percent composition 

 data are statistically very reliable for both rare and common pollen 

 grains. If the count includes 200 or fewer grains, the calculated per- 

 cent composition involves a sampling error that becomes increasingly 

 serious for progressively smaller counts, and is more serious for com- 

 mon pollen types in the sample than it is for rare ones. When two 

 or more pollen diagrams from a deposit are essentially the same, 

 reliability of the data is increased (Faegri and Iversen, 1950). 



The pollen diagram shown in figure 6 represents an analysis of a 

 4-meter core in a late-glacial bog near Totoket Mountain, Northford, 

 Conn. As a help in visualizing the sequence, a diagrammatic section 

 of the core sediments from which the pollen samples were taken is 

 shown at the left of this figure along with a scale to show depth below 

 the surface of the bog. Plant genera observed in the samples are ar- 

 ranged from left to right starting with trees, followed by shrubs, 

 herbs, and water plants. The relative amounts of different pollen 

 types are shown as deviations to the right from the vertical axes. 



The advantage of presenting data in a diagram is that at a glance 

 one can easily see major trends, dominant types, and relations of 

 sediment type to the pollen phases, and also observe the components 



