Let P(t) be the largest prime divisor of a positive integer t>1. Let m be the largest odd divisor of n:n=2km. Then 2n+1=22km+1=am+1, where a=22k. If k>0, that is, n is even, then C(n)=C(m)+2 and C(2n+1)=C(am+1). We need the following two lemmas. Lemma 1. For every prime p>2 we have P(a+1ap+1)=p or P(a+1ap+1)⩾2p+1. Proof. Let P(a+1ap+1)=q. It follows from Fermat's little theorem that q divides 2q−1−1 and therefore (a2p−1,aq−1−1)=a(2p,q−1)−1. The greatest common divisor (2p,q−1) is even and must equal 2p or 2. In the first case 2p divides q−1, whence q⩾2p+1. In the second case q divides a2−1 but not a−1 (because ap+1 is divisible by q), that is, a≡−1(modq). Then a+1ap+1=ap−1−…+1≡p(modq) and p=q. Lemma 2. If p1 and p2 are different odd primes then P(a+1ap1+1)=P(a+1ap2+1). Proof. If P(a+1ap1+1)=P(a+1ap2+1)=q then q divides a2p1−1 and a2p2−1, therefore (a2p1−1,a2p2−1)=a(2p1,2p2)−1=a2−1 and hence a+1, but then p1=q and p2=q, a contradiction. We are ready now to solve the problem. Let p1,…,ps be all the prime divisors of n. It follows from Lemma 2 that C(2n+1)⩾P(a+1ap1+1)+…+P(a+1aps+1) If C(2n+1)>P(a+1ap1+1)+…+P(a+1aps+1), then 2n+1 has at least one prime divisor not summed in the L.H.S., that is, C(2n+1)⩾P(a+1ap1+1)+…+P(a+1aps+1)+3⩾p1+…+ps+3>C(n) Therefore we can assume the equality: C(2n+1)=P(a+1ap1+1)+…+P(a+1aps+1) If in this case there is an i such that P(a+1api+1)=pi, then C(2n+1)⩾p1+…+ps+pi+1>C(n). It remains to consider the case when P(a+1api+1)=pi for all i. In this case we have C(n)=C(2n+1)=p1+…+ps, so n must be odd and a=2. But 2p≡2(modp) for all odd prime p, therefore p>3 cannot divide 2p+1. Thus s=1,p=3,n=3r with some positive integral r. The number 2n+1=23r+1 must be a power of 3. However 19 divides this number for r=2 and consequently for all r⩾2. Thus the only remaining case is n=3, which obviously satisfies the condition.