The key concept here is Min-Max optimisation under integer constraints — and this question has two separate sub-problems: maximise the total coffees sold, and minimise it.
Setup: 2S + 3M + 4L = 404, with M > S >= 1, M > L >= 1, all integers.
--- PART 1: Maximise total (S + M + L) ---
To maximise count, we want as many cheap coffees as possible. Small ($2) is cheapest, but M must be the largest group. Strategy: set L = 1 (minimum), then maximise S while keeping M > S.
With L = 1: 2S + 3M = 400, and M > S.
From 3M = 400 - 2S → M = (400 - 2S)/3.
For M to be an integer: 400 - 2S must be divisible by 3. Since 400 ≡ 1 (mod 3), we need 2S ≡ 1 (mod 3), so S ≡ 2 (mod 3).
For M > S: (400 - 2S)/3 > S → 400 > 5S → S < 80.
Largest valid S: S = 77 (since 77 ≡ 2 mod 3 ✓ and 77 < 80 ✓).
→ M = (400 - 154)/3 = 246/3 = 82.
Check: M = 82 > S = 77 ✓, M = 82 > L = 1 ✓.
Max total = 77 + 82 + 1 = 160.
--- PART 2: Minimise total (S + M + L) ---
To minimise count, use the most expensive coffees. Large ($4) is priciest, but M must still be the largest group. Strategy: set S = 1 (minimum), then maximise L while keeping M > L.
With S = 1: 3M + 4L = 402, and M > L.
From 3M = 402 - 4L → M = (402 - 4L)/3.
For M integer: 402 - 4L divisible by 3. Since 402 ≡ 0 (mod 3) and 4L ≡ L (mod 3), need L ≡ 0 (mod 3).
For M > L: (402 - 4L)/3 > L → 402 > 7L → L < 57.4.
Largest valid L: L = 57 (since 57 ≡ 0 mod 3 ✓ and 57 < 57.4 ✓).
→ M = (402 - 228)/3 = 174/3 = 58.
Check: M = 58 > L = 57 ✓, M = 58 > S = 1 ✓.
Min total = 1 + 58 + 57 = 116.
--- ANSWER ---
Difference = 160 - 116 = 44. Answer: C.
---
Common trap (dec2023's approach above): Setting M = 128 violates the problem constraints — you also need M > S, so S can't be as small as 1 when M = 128 unless S < 128. But that approach also doesn't correctly maximise the total number of coffees. The goal is to maximise S + M + L, not just maximise M.
Takeaway: On Min-Max word problems with three variables, always anchor one variable at its extreme (min or max), then use the divisibility constraint to find the binding limit on the other.
(Kavya | 725 on GMAT Focus Edition)