I have two ancestors that I cannot find in 1870: John Ufkes and Ira Sargent.
Both men were single in 1870, under thirty-five, had never been married, and had no children. Ufkes probably lived in the general area of Hancock County, Illinois, and Sargent probably in the area of Union County, Iowa. Both were likely working as hired men or laborers of some sort. That's part of the problem--they were unattached and had relatively mobile occupations.
But the question remains: "When have I looked enough?" At what point do I decide the time searching is not time well spent and is time I could devote to other research activities?
It is a question that does not have a hard and fast answer and really depends upon the individual in question. It also depends upon my personal research philosophy.
Ufkes is well documented and has been located in a variety of records in the United States. He's generated and left behind virtually every record a German immigrant farmer who lives in the United States from 1869-1924 should leave--except for his 1870 census enumeration. His place of birth in Germany is known, as are the names of his siblings, parents, and extended family. I've conducted a reasonable search for him in 1870 and don't think the enumeration of him will probably alter the life chronology and portrait I already have for him. It would be personally satisfying to find him, but it's not likely that the enumeration will provide me with an entirely new avenue of research on Ufkes (possible, yes--probable, no). Because of that, I have stopped looking for him 1870. Would I like to know where he is? Yes. Is it worth manually searching three states to potentially locate him? Probably not.
Sargent is not as well documented and comes from a hazier background than does Ufkes. Sargent's family was moving around Iowa and Missouri in the 1860s. His parents names are known, but not much is known about his mother's family. He marries in Union County, Iowa, in October of 1870, so it seems reasonable that he's there at the time of the census, although it is possible that he is not. Most of Ira's siblings have been located in 1870 (he's not with or near them) and it's possible that he's living with one of those who cannot be located. Ira's 1870 enumeration has a greater chance of assisting me with my research on him than Ufkes' does. If Ira could be located, his residence could provide another location in which to conduct a general record search.
Both men were probably working as hired men or laborers in 1870, perhaps living with a family and enumerated with that family's last name. They could have been in between residences at the time of the census and simply not enumerated. As unattached men, they simply could have been missed by the census enumerator.
Sometimes I occasionally look for Sargent--just to see if I overlooked something.
While I like to determine as much about each ancestor as possible, there comes a time when one has to admit that there may be other ways to spend your genealogical time.
After all, there are other ancestors whose stories are just begging to be researched as well.