Find_Housing wrote: > "David Heiser" <[EMAIL PROTECTED]> wrote in message news:<[EMAIL PROTECTED]>... > >>Findamental is how well is the medium mixed. 10 bacteria makes sense if the >>medium is truely homogeneous in all other constituents. The problem then is >>the classic "balls-in-an-urn", since a bacterium is a finite particle, and >>everything else is not a particle. If the medium is a gel, then the issue is >>primarily one of homogenaity, and you have to take many more samples. One >>then gets involved in a designed experiment involing three dimensional >>sample locations. >> >>David Heiser > > > Thank you David. It's liquid, not gel. > I am trying to readress the question as follows: > > The question is now "how to distinquish a bag that is 1000 bacterial > per ml from a bag that is 10 bacteria per ml? How many samples of 1 ml > should be taken out of a 60 ml bag? (0.9 power and 0.05 level of > significance) > Any way to estimate the variance? > > Aron
I'm not sure that's your question for a few reasons. Your real question seems to be about the actual bacterial concentration level in the bag, not that it's 10 or 1000. If it's a liquid and the bacteria don't clump (or else see Student's paper on haemacytometers--Bk, 5, 351-360), your ability to estimate the concentration will be determined by the uncertainty in your measuring instrument, that is, the measuring instrument will determine how many samples you need to determine the concentration to a particular precision. It would like testing for something in your blood. We rarely require a pint because we might miss the virus in a particular tube. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
