250 Forestry Quarterly 



subjective conceptions as regards the frequency of plants, by 

 laying out a number of quadrats within which all plants are 

 counted and thus their "valency" mathematically determined. 



The question arises how large the squares and how many must 

 be investigated to arrive at a true valency of species. Theoreti- 

 cally, squares of smallest extent and a large number of them 

 would give the surest results ; practically, this is not attainable. 



For the case in point — investigating floral changes on sample 

 areas of one half to more than one acre — a systematic location 

 of quadrats appeared desirable. Not only the relative frequency 

 of occurrence, but the area covered by species, especially those 

 which are of significance for humus formations (Vaccinium, 

 etc.), it is desirable to know, hence, besides a frequency per cent, 

 also an area per cent must be secured ; this, to be sure, by estimate 

 in each square or quarter square, as the author prefers to reduce 

 the subjective element. As an expression of vigorousness of 

 development, the maximum height of the species was also noted. 



To determine what size area would give best results as regards 

 species distribution, a brass frame with movable sides was con- 

 structed by which Vs. V4. Vs. down to Vio sq. m. could be en- 

 closed. Squares were laid out, spacing them 2 by 2, 4 by 4, and 

 8 by 8 w in systematic manner. Since frequency and area per 

 cent, as well as average height, are not exact figures, the average 

 error was then established by the method of least squares. 



Altogether seven sample areas were investigated, 16 tables and 

 four diagrams showing the results. 



The maximum error, if quadrats are laid out with a spacing of 

 4 by 4 m, lies usually between ± 5 to 6 per cent, rarely d= 10 

 per cent; with a spacing of 2 by 2 m, under ± 2 per cent, but 

 with 8 by 8 m as high as 20 per cent. The area percentage can 

 therefore be very exactly ascertained with a 2 by 2 spacing, and 

 still satisfactorily with a 4 by 4 spacing. The size of the squares 

 used was .5 m and .1 m. No attempts seem to have been made to 

 ascertain which size furnishes best results. A short bibliography 

 is added. 



Markflorans analys pd objectiv grund. Meddelanden frlin Statens Skogs- 

 fdrsoksanstalt, 1914, pp. 129-200. 



