^{Note (1) page 54} G. F. Ray, ‘The diffusion of new technology: a study of ten processes in nine industries’, National Institute Economic Review, no. 48, May 1969.

^{Note (2) page 54} R. J. Smith, ‘The weaving of cotton and allied textiles in Great Britain: an industry study with special reference to the diffusion of the shuttleless loom’, National Institute Economic Review, no. 53, August 1970.

^{Note (3) page 54} Fletton bricks are made from a special type of clay (lower Oxford) found only in this country—in a belt running across the Midlands and centred mainly in the Bedford-Peter- borough area. Because its high carbon content provides most of the fuel needed, substantial economies in burning are possible; and its high moisture content means that no water needs to be added before the bricks are shaped and the drying stage is totally dispensed with. Fletton and non-fletton bricks are virtually interchangeable.

^{Note (4) page 54} There is no precise definition of industrialisation. According to the Ministry of Housing and Local Government (circular no. 76/65) it covers ‘all measures needed to enable the industry to work more like a factory industry’.

^{Note (1) page 55} The firms to which the questionnaire was sent included all the large and medium-sized firms and most of the smaller ones currently producing clay bricks. The Institute is grateful to the 85 firms which gave useful replies. These represented a response rate of at least 65 per cent in terms of numbers of producers and 86 per cent in terms of output (100 per cent for flettons and 74 per cent for the rest). The true rate would be higher to the extent that the non-respondents included firms which had recently ceased clay brick production. It was, however, lower for individual questions, notably on invest ment, and the sample is biased in the direction of the larger firms.

^{Note (2) page 55} This account of the 19th century brick industry is based largely on M. Bowley, Innovations in building materials, 1960 and H. W. H. West, The establishment of the brick and tile industry in developing countries, UNIDO, 1969.

^{Note (3) page 55} The ‘wire-cut’ or ‘extrusion’ process, used for the more plastic clays, accounts for a large percentage of continental and British non-fletton production. It requires water to be added to the ground clay, which is then pushed through a mouthpiece over a cutting-table in a continuous column that is cut, by wires, into bricks of the required thickness. In the ‘semi-dry press' process, used predominantly by the fletton sector (footnote (3) on page 54), the clay is shaped under pressure in closed moulds without the addition of water.

^{Note (4) page 55} The principle of the ‘continuous' kiln is that the firing process is uninterrupted (except for maintenance). In the ‘tunnel’ type the fire is maintained continuously in one part of a long tunnel and the bricks, placed on waggonettes or cars, pass through the firing zone at a uniform speed, whereas in other types the bricks are stationary while the fire moves.

^{Note (1) page 56} NBPI Report no. 47.

^{Note (2) page 56} NBPI Report no. 150.

^{Note (3) page 56} No accurate figures are available, but the chairman of National Star Brick and Tile Holdings Ltd (one of the leading non-fletton producers) stated on 20 May 1970 that in the three years ending 31 December 1969 107 works were closed, 50 in 1969 alone.

^{Note (1) page 57} An industry expert estimates that the number of brick making firms of all sizes in 1971 was about 370. Apart from economic factors, exhaustion of clay seams has accounted for the disappearance of some firms.

^{Note (2) page 57} Unlike LBC, which is almost totally concerned with brick production, only eight of the firms have more than 50 per cent of their turnover in brickmaking.

^{Note (1) page 58} The Cement Makers Federation estimates that 20 per cent of cement deliveries derive directly from housebuilding, although another 20 per cent are used in concrete products (more than three fifths of which are used in housebuilding) and a further 12 per cent go to builders' merchants. In all, therefore, up to about 45 per cent of cement deliveries may be attributable to housebuilding.

^{Note (1) page 62} Report no. 47.

^{Note (2) page 62} The equation, which because of the smaller number of observations and the crudity of two of the variables is far from conclusive, is:

S/S = 4.355-0.618 P/P + 0.028 SL—10.279 B R² = 0.812 (0.281) (2.296) (2.404) DW=1.707 where S is the fletton's share of the market, P is its price in relation to the non-fletton price, and B and SL are dummy variables taking the value of 1 in periods of boom and slump respectively and otherwise of 0.

^{Note (3) page 62} Ex-works prices are used for non-flettons and delivered prices for flettons (taken from Trade and Industry), the latter being the effective price facing the non-fletton producers.

^{Note (4) page 62} Chart 5 above does indicate a slight improvement in market share in 1969 even though demand was falling.

^{Note (1) page 63} Regular data on manual workers' weekly earnings are available for ‘bricks and fireclay goods' but not for brick alone. Brickmaking, however, represents about 60 per cent of the whole category and Census data show hardly any difference in wages and salaries between brickmaking and the rest of the category in either 1954, 1963, or 1968.

^{Note (2) page 63} Apart from the possible influence of more capital-intensive methods in the big firms, this is, however, a very unreliable measure of productivity. It ignores, for example, any change in the facings/commons ‘mix’, which would affect both factor inputs and the value of output.

^{Note (1) page 64} In the analysis in this section it is impracticable to divide non-users of the TK between users of other types of con tinuous kiln and users of intermittent types. It is, however, important to remember that the non-users include both these categories and that the differences would be smaller if the comparisons were between TKs and other continuous kilns only. This applies particularly to table 11.

^{Note (1) page 66} In fact, this is an underestimate to the extent that the 1969 output of the TKs was below capacity.

^{Note (2) page 66} Mansfield (Industrial Research and Technological Innovation, New York, 1968), for instance, postulates that a firm's size and its liquidity will both affect the rate at which it turns over its output to a new innovation. Nelson and Phelps (American Economic Review, Papers and Proceedings, 1966) postulate that the educational attainment of managers may affect the rate at which they adopt a new technology.

^{Note (1) page 67} Whether measured by firms having introduced at least one TK or by total numbers of TKs or output from TKs.

^{Note (2) page 67} Since it meant that future discounted returns need not exceed the cost of dismantling any present kiln or any con tinued returns from that kiln.

^{Note (1) page 68} Data are only available for firms' growth over the period as a whole and most of them adopted at some time between 1960 and 1969.

^{Note (2) page 68} In this case, because of the extreme durability of the kiln, he curve is likely to be stretched out.

^{Note (1) page 69} That is the point of inflexion of the curve (given by t = -B/2).

^{Note (2) page 69} Diffusion will asymptotically approach A.

^{Note (3) page 69} Because separate figures are available for common bricks only.

^{Note (4) page 69} More rigorously, various Almon distributed lags were tried, but still the coefficients of Pt, Pt-i, and P_t-2 remained positive and insignificant. The best R² was obtained by using a polynomial of degree three.

^{Note (1) page 70} Where labour costs were represented by an index from Monthly Bulletin of Construction Statistics of the weekly earnings of manual workers in bricks and fireclay goods production.

^{Note (2) page 70} Technically speaking, TSLS yields consistent estimates but does not necessarily remove bias; however, as the number of observations here is quite large (60) there is some chance of ‘better’ coefhcients than those derived from single equation estimation.

^{Note (1) page 71} In every case a log linear specification performed slightly better than linear ones.