Dear Listservers
I would like to discuss in this message the detection of ammonia and TMA in
fish. The thoughts were stimulated by Andrew Strak's comment in his message
of May 17 that the detection thresholds of ammonia and TMA in fish are 110
and 0.6 ppm respectively. If Andrew is reading this I would be grateful for
his source of these data. A compilation of odour threshold values of
chemicals in water I have does not list a value for ammonia and lists two
values for TMA - 0.5 and 1.7 mg/l (ppm). I have a paper by Amoore &
Forrester, 'Specific anosmia to trimethylamine: the fishy primary odor,
Journal of Chemical Ecology, 1976, 2, 49-56, not cited in the compilation,
which gives 23.9 and 0.000467 ppm for detection thresholds in water of
ammonia and TMA respectively. It is quite common in compilations of
detection thresholds to see values for a chemical having over a fold range
of concentrations, and a thousand fold is not uncommon, so the Andrews and
the Amoore & Forrester values for ammonia are not different in my view. The
difference in values for TMA, around thousand fold, is unusual, but not
unprecedented. There are several reasons why threshold values should vary
among determinations in different laboratories, chief among them are
differences in experimental procedures for presenting the stimuli to the
assessor, in criteria for determining detection, selection and training of
assessors, and definitions of and calculations of the thresholds. There are
further particular problems in determining thresholds of ionising substances
like amines.
Chemical analysis will measure, for example, total ammonia, but in solution,
including fish tissue, part of the ammonia will exist as free base and part
as the ammonium ion. It is the free base that is detected as odour. The
proportions of each form is dependent on the pH of the medium and can be
calculated from a knowledge of the dissociation constant, Kb. Similarly for
TMA. Ammonia and TMA are similar in their strengths as bases and their Kb's
are 1.74 x 10 to the -5 and 7.4 x 10 to the -5 respectively. The proportions
of total ammonia as the free base are 0.003, 0.004, 0.005 and 0.006 at pH's
of 6.72, 6.85, 6.95, and 7.02 respectively. For TMA the values are 0.004,
0.005, 0.006, 0.007 and 0.008 at PH's 6.73, 6.83, 6.91, 6.98 and 7.04. Less
than 1% of the total chemical specie at pH's of even spoiling fish muscle.
The ammonia content of non-elasmobranch, vertebrate, fish muscle is around 8
mg/100g by the time of resolution of rigor mortis - it comes from
deamination of adenosine to inosine - and increases very little during
spoilage. In cod it about 11mg/100g at 20 days in ice. Presence of ammonia
is not an indication of spoilage in non-elasmobranch fish. If the proportion
of free base is 1% of this its concentration is 0.11mg/100g or 1.1 ppm. This
is well below the reported detection thresholds of ammonia. Elasmobranchs
contain about 2% urea in the flesh. On extensive spoilage most of this is
converted to ammonia, equivalent to around 0.6 g ammonia/100g. Assuming 1%
is present in the base form gives a concentration of 6mg/100g, 60 ppm. This
is above the detection threshold of ammonia and spoiled elasmobranch flesh
smells of ammonia.
TMA content of unspoiled fish is very low, less than 0.5mg TMA-N/100g, (2.3
mgTMA/100g), and typically increases to around 15 mg TMA-N/100 (70
mgTMA/100g) in spoiled fish at around 15 days in ice. Again assuming 1% as
the free base gives a concentration of 0.07mg TMA/100g or 0.7 ppm. This is
at, or well above, the reported detection threshold.
(Would someone check my arithmetic).
If sensory score sheets include 'ammonia' as a descriptor of spoiled fish,
then it is TMA that is responsible, not ammonia. Both have similar odours,
but they are not identical.
I regret now that I am retired and have time to think about research rather
than doing it, that I did not investigate the sensory properties, including
detection thresholds, of ammonia and TMA. I commend it as a topic for a
student looking for a research project. It will not be as easy at it
appears. Accurate and precise measurements of thresholds is not easy. Do not
be taken in by the typical descriptions in textbooks of sensory testing.
Peter Howgate
This archive was generated by hypermail 2b29 : Sun May 21 2000 - 14:51:30 PDT