Samples of PAN were treated separately in vacuum and in atmospheres of ammonia and air，at each of the temperature，200°，250°，300°，350°，and 400℃，one after an-other for a definite length of time．Before and after each step of treatment，the weight of the sample，its infra-red spectrum，its solubilities in cone．sulphuric acid and in glacial acetic acid as well as the ultra—violet spectrum of its solution in cone．sulphuric acid were examined．The conditions of treatment are given in Table 1，and the results，in Figs．1—4．From a careful examination of these measurements

it was discovered that different schemes of mechanism have to be assigned for the reactions taking place in the different atmospheres and also at different temperatures．At lower temperatures，the first stage reactions taking place in vacuum(200°) and in atmospheres of ammonia (≤180°) and air (200℃) are considered to be the formation of the respective structure units：(Vlb)，(N1) and (O1)．After this，cyclization follows and it will proceed with tIle formation of (V2)，(N2) and (O2) one after another．But as soon as any two of the following groups：C—X，C=NH，C—NH2，C=O should come to meet at neighbouring positions，cyclization will be immediately stopped． Moreover，the different orientations of the CN groups with each other in atactic samples will also prevent cyclization to proceed further．With the increase of temperature，the rate of cyclization may be accelerated，and be-sides，additional cyclization may be obtained by the splitting of NH3 from the neighbour-ing C=NH，C-NH2 groups．But here again，the increased cyclization will be limited by the C—X and C=O groups．and by the atactic nature of the sample．Each of the molecules thus obtained will still be composed of a series of small units of polynuclear pyridine rings of different lengths and orientations，irregularly scattered along a carbon chain，[(V3) and (O3)]． The schemes of reaction mechanism suggested above agree nicely with all of the chief characteristics of our experimental facts given in Figs．1—4．and also with the low values of semiconductivity of the products。which are generally of the order of 10-9—10-7 Q-1 CM-1，obtained from measurements at 20°一300℃．From these

it will be seen that no large extended polynuclear pyridine systems of uniform composition and regular struc-ture could be formed as would be expected from the simple theories suggested by pre-vious authors．And hence the explanation made by some writers for the semiconductive properties of the products based on the formation of such an ideal structure is considered to be inadequate．