You don't give much information upon which to make a guess.  If you can provide the following info, it might help.

1.  What is the magnitude of the difference?

2.  Do you do duplicate blanks, and if so, is there a similar difference between the two blank results?

3.  In what sequence do you set up and read the bottles in a batch?

4.  What is the relationship between the first and second GGA bottles with regard to location in the incubator?

To answer the questions, 1) the first bottle will often have about a 0.3 to 0.8 mg/l higher depletion.  

2) I don't incubate one of the 2 blanks. I just use it to get an initial DO.

3) Order is  blanks, standards, seed , final,  raw, others

4) All bottles are in a 12 bottle carrier. the standards are the 2nd and 3rd in back corner after the blank. The incubator is the size of a 1/2 size refrig. (like undercabinet)

Hmmm...

I asked about the sequence of bottles in the batch to see if perhaps you were doing the seed control just before the GGA, which could lead to carryover...but that's not the case.

You might try switching the order of the GGA bottles in the 12-bottle carrier.  One of the two bottles might be a tad warmer than the other because of proximity to a source of heat in the incubator (a converted fridge?).  If the second bottle then consistently has higher results than the first, the permanent solution might be to put a small fan in the incubator to circulate the air, to place the entire carrier in a different location in the incubator (if possible), or to replace the incubator.

One other thing I should have asked is whether you use a DO meter/probe, or Winkler titration to measure DO.  Unless you are doing Winkler's, there is no need to take an initial DO on one blank bottle and incubate the other one.  If you are using a meter/probe, you might try measuring both initial and 5-day DOs on both bottles and see if you experience one bottle always being a little higher than the other.  Alternatively, you could do a few effluent samples in duplicate.  In the long run, however, you don't get much useful information from doing blanks in duplicate, and the same can be said for doing effluent in duplicate if your precision is good (as indicated by a low standard deviation on GGA results.

There might be some other cause, but I sure can't think of what it might be.

Hi Tom:

Are the GGA samples the first bottles filled?  I've seen cases where the dilution water tubing gets a growth on it.  The initial samples got contaminated as some of this growth was rubbed off during the first bottle filling.  Periodic bleaching eliminated the issue.

Dean