Interesting.

I do think that calling it milk is silly, however.   Milk is universally defined as a secretion of mammary glands in mammals.   To define other secretions as such undermines standard definitions (which is unwise).   So, while I take the point that "Nutritive fluid would be more general" it doesn't "tell the same" at all.

Interesting, though.

The peak at 230nm is likely to be not from your DNA (phenol and carbohydrates typically absorb at this wavelength, Trizol gives a response at 230nm and 270nm).
You say 2ng/ul is "rubbish" but neglect to tell us how much DNA should be extracted and what the elution volume is.  

You might want to do a 20mg/ml proteinase K, 1M DTT digest and then spin the extract through a suitably spinfilter to concentrate the extract.   But put in the amount of intact hair bulbs to give you the amount of DNA you need (whatever that is).

Milk production in invertebrates?   Unless I'm missing something I think someone's playing a joke.

A couple of things:

If you're going to use Durans then only use them.   Don't use Erlins for some ferments and not others.   That way you will have consistency of results and won't introduce artefacts in your results.

If you seal the Durans then that'll keep oxygen out but how will you vent CO2 produced?   Given everything is sterile then a cotton bung and aluminium foil will allow the positive pressure of the CO2 created to vent and create a positive pressure environment to help assure that nothing gets in...if everything is sterile to begin with.

It depends upon several factors - the chemical you're using and the use you have for it.

I'll take the second point first.   If you are using the chemicals for academia then 'all' you are jeopardizing is your work.   If you are working in a regulated environment then you are jeopardizing your Quality Assurance system.

As for how stable the chemicals actually are, well, again that depends.   Expiry dates are founded on the results of stability data.   Stability trials (either accelerated or real time) have an endpoint and there is no company that produces these chemicals that will run endless trials so that their expiry dates are more than 5 years at the very most.   Shorter expiry dates may be because of inherent instability (i.e. stability trials failed for some samples at a certain point) or expense in running trials (the chemicals may inherently be expensive to handle so stability trials may end after a year or whatever time length).

Using chemicals within expiry also is a partial safeguard against contamination.   For example, if you have a tub of sodium chloride, I'd put a year expiry on that.   Why?   Yes, that stuff is going to be stable until the Sun goes nova but if you have it kicking around for 3, 5, 10, 15 years (I once inherited a tub of urea that was more than 15 years past it's expiry and the container that originally held it disintegrated) then do you really think that that hasn't been contaminated in that time?   The material is so inexpensive that you really don't want to ruin even a small experiment for the want of decent grade chemicals.

So, it's a balance - how long is it stable, how sure you are that it's uncontaminated, how expensive it is to replace, how evident that it's gone 'off' (restriction enzymes are a good example - it's really easy to check them and they are stable for years if stored well), how disastrous is it to your experiment/lab/reputation/product if you get it wrong...et cetera.

Learn to identify when specific chemicals have degraded (colour, smell, separation, viscosity, et cetera) and always check for contamination as best you can (e.g. homogeneity of appearance) and even if you are in academia think about having a system of replacing critical materials on a scheduled basis.   It's just good practice.