MICROPALÆONTOLOGY OF POSTGLACIAL DEPOSITS, 453 
The PF (frequency of pollen per preparation ; 560 in the above record) 
indicates the number of pollen-grains in a unit of area (N.B. The pollen of 
forest trees ; not Corylus, Saliv, Rhamnus, ete.). The unit I have chosen is 
l sq. cem. PF is by no means an exact number, because the thickness of the 
preparations may vary considerably. After practice, however, the prepara- 
tions can be made, as a rule, of almost uniform thickness, in which case the 
PF numbers may have a certain importance. 
To get almost exact percentages it 1s sufficient to count about. 150 pollen- 
grains; but even 1f only 50 pollen-grains are counted, the result is of 
value (vide Jessen, 1920, p. 23, note). Auer (1923, pp. 347-349), in his 
explorations of peat in northern Finland, expressed the PF by using a scale 
with 5 degrees (1— very rare, 5= very common). 
In the records of analysis, after the PF numbers, I write, in brackets, the 
number by which the total of pollen-grains counted is to be multiplied so as 
to get the number 100 ; or, expressed in another way, the figure by which 
the number of pollen-grains of each species is to be multiplied so as to get 
the percentages of each species: for example, in one record we find 
PF=33 (2). There are here 33 pollen-grains per sq. cm., and the analysis 
is not quite reliable because only 50 pollen-grains were counted. In a 
second, PF —126 (1) ; here 100 pollen-grains were counted. In a third, 
PF is 88 (0°65) ; and thus about 155 pollen-grains were counted. As to the 
numbers in brackets, experience tells that 0—1 designates good, almost exaet, 
analyses ; 1-2 fairly good; 2-5 less good; and numbers >5 analyses in 
which the percentage figures are of little or no value. 
The relative frequency-number of Corylus and Salix, also of other micro- 
fossils whose frequency one desires to express (e. g. Nuphar pollen, tetrads 
of Erieaees, spores, etc.), is calculated separately as a percentage of the 
total number of pollen-grains of the forest trees. Owing to the fact that 
Salix pollen somewhat resembles the pollen of certain plants with a very 
different systematic position, I decided not to count the pollen of Saliz with 
the pollen of. the forest trees as von Post and others had done previously and 
I myself did in 1921. , 
In Scotland I have found fossil pollen from all trees, Picea exeepted, of 
which the pollen is met with in the peat mosses of Southern Sweden : Acer, 
Alnus, Betula, Carpinus, Fagus, Fravinus, Pinus, Quercus, Tilia, and Ulmus ; 
also in a few localities I have discovered //ew pollen. 
By means of the percentage numbers a pollen-diagram is constructed, 
* The relative frequency-numbers, which are produced for the pollen-species 
found in a sample investigated, constitute the pollen-speetrum of the sample. 
On the basis of a series of * pollen-spectra ^ from a sectional boring in a bog, 
a pollen-diagram may be constructed, in which curves for the single species, 
or for a group of species, give both a visual representation of the composition 
of the pollen-flora and the oscillations as regards frequency which have 
taken place reciprocally between the pollen-curves during the formation 
