(a) Take two readings x1,x2∈R that produce the same score:
3x1−4=3x2−4.
Add 4 to both sides: 3x1=3x2. Divide by 3 (which is allowed, since 3=0): x1=x2.
So two different readings can never give the same score, and f is one-one. In plant terms, a score determines its reading unambiguously — which is what makes part (c) possible at all.
(b) Let y be any real number, and ask which reading would produce it. Solve 3x−4=y:
3x=y+4,x=3y+4.
This x is a genuine real number for every real y, so it is a legal reading. Substituting it back:
f(3y+4)=3⋅3y+4−4=(y+4)−4=y.
Every real score is therefore attained, so f is onto R. Together with part (a), f is a bijection from R to R.
The graph says the same thing: the line y=3x−4 has slope 3=0, so every horizontal level meets it, and meets it only once.

(c) Because f is a bijection it is invertible, and part (b) has already done the work of solving for the reading. The supervisor should use
f−1(y)=3y+4,
where y is the logged score and f−1(y) is the raw reading in millivolts.
(d) Put y=11 into the formula from part (c):
f−1(11)=311+4=315=5.
The raw reading was 5 millivolts. Check it forwards: f(5)=3×5−4=15−4=11, which is the logged score. The answer is confirmed.
(e) h is one-one but not onto.
One-one. The algebra of part (a) works just as well inside Z: if 3x1−4=3x2−4 with x1,x2∈Z, then 3x1=3x2 and so x1=x2.
Not onto. Take the score y=1, which is an integer and so belongs to the codomain Z. A pre-image would have to satisfy 3x−4=1, that is x=35, which is not an integer. So 1 is a score the older rig can never record, and h is not onto.
More generally h(x)=3x−4 leaves a remainder of 2 on division by 3 for every integer x, so the only achievable scores are …,−7,−4,−1,2,5,8,… — exactly one integer in every three.
This is worth pausing on. If f mapped a finite set to itself, one-one would have forced onto. On an infinite set such as Z that implication collapses, and h is a plain counterexample.