Some incidental information about the JONES family
In 1881 John Jones (30), an agricultural labourer born in Montgomeryshire, was living at Pentre, Llanfair Caereinion, with his wife Elizabeth (25), daughter Mary A Jones (I believe this was the Mary Ann Jones who, together with her husband eventually adopted and raised Gwladys Mary Jones), and Elizabeth's mother, Margaret Jones (72).
By 1891 they had moved to Brynrefail (various spellings in records over the years), Llanfair Caereinion, and the household was John (41), a general labourer born in Llanfair, Elizabeth (35), hereafter listed as born in Meifod, John Edward Jones (7) (he married Mary Brown - aka Mary Starkey - see previous page), Esther (3) and David Richard Jones (1); David died the following year.
Thereafter John (sen.) was not with the family in the censuses but Elizabeth continued to assert she was married.
In 1901 Elizabeth (45) was still at Brynrefail with Esther (12), Elizabeth Jane (4) (I believe she became the "Auntie Jinny" to John Edward Jones's children Gilbert and Gwladys) and Gertrude (7 months).
In 1911 Elizabeth (54), Elizabeth Jane (14) and Gertrude (10) were still at Brynrefail. Interestingly, it stated Elizabeth had been married 32 years and 5 of her 10 children had survived. All this was crossed through, presumably because of the absence of her husband. The 4 other missing children must have been born and died between census years.
Note that all the above were born at Llanfair Caereinion unless stated otherwise.
In 1901, I believe John Edward Jones may have been a cowman with Thomas Jones at Neuodducha, Llanfair, and there was an Annie Jones (13), born in Llanfair Caereinion, living as granddaughter with John and Sarah Jones at Upper Trefnant, Guilsfield; this might have been Mary Ann who was known as Annie.
In 1939 the only one of the above living at Brynrefail was Elizabeth J Jones ("Auntie Jinny") (b 1 July 1896) with someone whose record was officially closed; usually such records related to young people.
Not an easy line to trace due to the fluidity of the family grouping so I hope this might be of some help to fellow researchers.