Bellifortis,
Midgley was a tremendous British engineer but lacks full recognition for his work. As you say his inventions include a car dynamo, car lighting systems, fuzes, a single (or was it 5?) valve radio, electronic organ, switches and timers for domestic appliances. Albert Midgely's son Bertie lived quite near to me and we became good friends. Bertie had worked with his father from about 1926 and had probably not thrown anything away from the day he joined the firm. Bertie had a pair of adjoining semi-detached houses just absolutely full of documents and artefacts (e.g. five 'W' bombs and prototypes). His family home was in Moor Park, a rather expensive area to live in. I can expand on this if required but in answer to your question:
'How did he get into the fuze business?' - Midgley was tied up with CA Vandervell Ltd working as their Chief Engineer during WW1. After the war he was awarded a sizeable sum for his invention of the 'allways' fuze (No 146) so he had a track-record for munitions design. Although by 1939 he had lost touch with the government departments dealing with munitions design by chance he met an acquaintance who was involved with the development of the 'W' Bomb. The result of this meeting was that Midgley worked with MD1 ('Winston Churchill's Toyshop') and effectively redesigned the 'W' Bomb for them. Midgley also designed a delay device for sabotage purposes. The relationship between Macrae at MD1 and Midgley was not a harmonious one and at some stage Midgley started to deal directly with the RAF's Director of Armament Development, particularly on fuze matters.
Midgley the person, and Midgley Harmer Ltd were different legal entities and Midgley goes to some length in his papers to say that his designs, which were drawn in his home-office in Moor Park, belonged to him. This would be a point of discussion after the war when the Royal Commission on Awards to Inventors considered his claims. M-H Ltd was still his company and it made many prototypes and did some production. I don't know how may staff there were.
The algorithm for decoding the patent number from the 'Application Number' involves Find printed copy of patent abridgements; mark pages involving Midgley; turn to page; look for N digit application number and date; repeat; be lucky!
Gelatine-H2O delay? Sounds an interesting idea but I don't know why he chose to mention it over an A-C system. McAlpines AC delay arrangement had been around since 1917. If I find anything I will let you know.
Answers are somewhat 'simplified'.