Emeline (Elliott) Cooper 1856–1910

Result: PARSED -> [G.697]. The bio field carries her SECOND HUSBAND'S full obituary from the Phillips County Review of 28 Feb 1928, which names her "Emeline Elliott Copper", dates that marriage to 3 Sep 1894 and her death to 2 Sep 1910, and lists her two Cooper daughters as his step-daughters. Also four Elliott siblings on the panel and a fifth, John M. Elliott, in the page's biographical note. THE FILE WENT FROM A BARE TREE-ONLY STUB TO A DOCUMENTED LIFE IN ONE COMMAND.

Result: The household did not surface under any of the five, and the null is a weak one for three reasons that are worth recording so the attempt is not simply repeated. (1) The index for these censuses is heavily initial-only - the Cooper and Mitchell result sets are full of "E Cooper", "S E Cooper", "M E Mitchell" - so anyone entered by initials cannot be reached by a given-name search at all. (2) The result list carries no county column. It shows census date, birth place, birth year, race and gender and nothing else, so a candidate can only be placed by opening its record, where "Residence County" finally appears. Triage therefore costs one page-load per candidate. (3) 1895 rows were a small minority in every one of the five searches, markedly rarer than 1885 rows. That is an impression from five result sets and not a measurement, and it is recorded as such - but if the 1895 index is thinner than the 1885 one, a null in it is worth correspondingly less. ONE CANDIDATE WAS OPENED AND REJECTED, recorded so it is not opened again: "Olive Mitchell, 1 Mar 1895, born Kansas abt 1885" is OSAGE COUNTY, locality Lyndon, family 136, in the household of W. C. Mitchell 43 and Maggie 42 with eight other children. Wrong county, wrong family, and the age fits neither of our sisters cleanly.

Result: No Cooper of any date is buried in Bow Creek Cemetery, and for once the null is a reasonably strong one. The control binds: surname Mitchell in the same cemetery returns exactly two, Emeline herself and Josiah, so the search mechanism works and the Cooper zero is a real zero rather than another silently unapplied filter. The pre-1900 sweep bounds it further: 39 burials, read in full across both pages. Every one carries a surname, so there is no anonymous "infant of" entry that a surname search would miss - and at least nine of the 39 are children under four (Dora May Andrews 1889-1889; Hubert Bartholomew 1886-1887; Mabel Bartholomew 1893-1893; Elsa, Hazel, Jasper and Leo DeLay 1893-1899; Minney Elliott Jan 1884, dead at eleven days; Clyde Horn 1896-1896; Guy Southard 1886-1886; Goldie Muir 1895-1898). This cemetery and its Find a Grave coverage plainly do record infant burials, which is what makes the absence of a Cooper infant worth something. What it does not close, said plainly: a grave can be unmarked and therefore unrecorded (89 per cent photographed describes photographs, not the completeness of the roster); a child could have been buried in Mitchell or Cloud County rather than at Glade; and any child of the Mitchell marriage after September 1894 would be a Mitchell, not a Cooper, and so outside this search entirely. By-catch, already relied upon and now attested on a stone: "Mary A. 'Molly Braide' Smith Mitchel, 1832 - 4 Jul 1886" (memorial 55429449) is Josiah's first wife, whose 4 July 1886 death date [D.874] uses in the arithmetic that excludes him as these daughters' father. That date had come from his obituary; it now has a gravestone behind it.

Result: NO COOPER OF ANY AGE IS INDEXED IN PHILLIPS COUNTY IN THE 1895 KANSAS STATE CENSUS, and for once the null is bound by a control that was run with byte-identical parameters: surname Mitchell returns 62 results of which 26 are Phillips County rows of this census, while surname Cooper returns none at all. The mechanism works; the Cooper zero is a real zero. THAT CONTROL WAS NOT A FORMALITY - the first two parameter shapes tried were both broken, and both failed SILENTLY. f.collectionId=1803994 combined with q.residencePlace returned "0 results" for Cooper AND for Mitchell, which would have read as a devastating null; and plain q.residencePlace without .exact is only a relevance hint, returning Cooper rows from Stafford County and rows with no county at all while displaying the Phillips filter. This is the fourth silently-unapplied filter met on this family. The rule that saved it is the cheap one: RUN THE CONTROL IN THE SAME SHAPE AS THE SEARCH. THE HOUSEHOLD ITSELF WAS THEN FOUND AND THE PLATE READ - see [E.55]. A Mitchel household of exactly three in RUSHVILLE TOWNSHIP, the same township [D.874] finds them in five years later: a man born "virgena" whose age is an overwritten blot (indexed 55), "M Mitchel" 28 born "ohio", and "R L Mitchel" 41 born "Kan". Neither Sadie (who would be 9) nor Olive (5) is in it, nor anywhere on the page.