(3) The idea that producers always operate in the range of diminishing 
returns. 
In analogy to Ricardo's picture of a fixed amount of land to which 
successive doses of input of labour and capital are applied, we have to picture 
a given amount of labour to which successive doses of capital are applied. 
This picture, however, would not be adequate to represent the conditions 
of an individual plant under modern industrial conditions, because technical 
progress may very well involve s reduction of the labour employed per plant. 
We plot, therefore, capital per man as a function of output per man. The gradient 
of a straight line joining the coordinate centre to a point of this function - 
the capital-coefficient - will decline over a certain range; later on we can assume 
for general reasons that it will rise again, or that it would if such methods were 
ever used. This increasing application of more units of capital per unit of 
labour is, however, constrained to a path of increasing scale of output capacity; 
A reasonable capital-intensification is in general not possible without an increase 
in output capacity: This is because the introduction of machines, apparatus, etc. 
immediately brings into sight dimensional economies which make the smaller scale 
inefficient. (It may nevertheless have to be chosen by firms without sufficient 
funds). Further, the use of continuous processes increases the technically 
possible utilization and, therefore, output scale (broad strip mill, continuous 
versus batch processes), 
This path of capital intensification can be regarded as a historical path-- 
perhaps we might say a historically necessary path -- because the development has 
really proceeded like that. Insofar as the techniques of the past are preserved 
in existing equipment, the path, however, is also mirrored to a degree in a cross- 
section of present plant, albeit with some modifications: In the course of evolution 
price-cost relations are brought down so that the capital-output ratio for the 
advanced methods is increased. At the same time the old-fashioned methods keep 
alive only by changing their output (special dimensions and qualities made in 
small batches, for example); this will prevent the output-capital relation from 
declining for the more old-fashioned firms (who could not exist if it would decline). 
Moreover, these series of historically successive techniques are also the 
ones which can be regarded as being simultaneously available to a prospective 
investor in the field. Whether we find in the range of these available methods 
all three stages (increasing, constant and diminishing returns) represented cannot