These processes may be broadly divided into three classes: 

  (1.) Processes which propose to combine nascent hydrogen with
  nitrogen at high temperatures or by electricity, with or without
  the presence of acid gases.

  (2.) Processes in which nitrides are first formed, from which
  ammonia is obtained by the action of hydrogen or steam.

  (3.) Processes in which cyanides are first formed and the ammonia
  obtained from these by the action of steam.

We began with an investigation of those processes in which a mixture of steam and nitrogen or of steam and air is made to act upon coke at a high temperature, sometimes in the presence of lime, baryta, or an alkali, sometimes in the presence of hydrochloric acid.

Very numerous patents have been taken out in this direction and there is no doubt that ammonia has been obtained by these processes by many inventors, but as I was aware that coke contains a considerable quantity of nitrogen, frequently as much as 1.5 per cent., which might be the source of the ammonia obtained, I determined to carry on the investigation in such a way as to make quite certain whether we obtained the ammonia from the coke or from the nitrogen of the atmosphere, or from both.  For this purpose we made for every experiment carried on by a mixture of nitrogen or air with steam another experiment with steam alone, carefully excluding nitrogen from the apparatus.  A very large number of experiments carried on at carefully determined temperatures, ranging from 500 deg. to 1,200 deg.C., and in which the directions given by the various inventors were most carefully observed, all led to the same result, viz., that the quantities of ammonia obtained were the same whether nitrogen was introduced into the apparatus with the steam or whether steam alone was used, thus proving conclusively that the ammonia obtained was derived from the nitrogen contained in the coke.

Further, on carefully determining the nitrogen in the coke used, it was found that the quantity of ammonia we had obtained in burning coke in a current of nitrogen and steam very nearly corresponded with the total nitrogen in the coke, so that we subsequently made our nitrogen determinations in the coke by simply burning it in a current of steam.

A process belonging to this class, proposed by Hugo Fleck, in which a mixture of carbonic oxide, steam, and nitrogen is made to pass over lime at a moderate red heat in order to obtain ammonia, was also carefully tried.  It was claimed for this process that it produced nascent hydrogen at temperatures at which the ammonia is not dissociated, and for this reason succeeded where others had failed.  We found that a considerable amount of hydrogen was obtained in this way at a temperature not exceeding 350 deg.C., and that the reaction was nearly complete at 500 deg.C.; but although we tried many experiments over a great range of temperatures, we never obtained a trace of ammonia by this process.