The problem of Hurwitz equivalence of n-tuples of braids appears in the study of braid monodromy of algebraic curves in C 2 . It was considered by many authors, see [1,3,4] and references in [3]. We give an answer for the case mentioned in the title.Let B 3 = A | σ 2 σ 1 = σ 1 σ 0 = σ 0 σ 2 , A = {σ 0 , σ 1 , σ 2 } be the Birman-Ko-Lee presentation (see [2]) of the group of braids with three strings. A quasipositive factorization of a braid X ∈ B 3 is a collection (X 1 , . . . , X k ) ∈ B k 3 such that X = X 1 X 2 . . . X k and for each i, the braid X i is conjugate to σ 1 . Note that σ 0 , σ 1 , and σ 2 are conjugate to each other. We denote the set of quasipositive factorizations ofThis action is called the Hurwitz action. Elements belonging to the same orbit are called Hurwitz-equivalent.It is proven in [4] that each orbit of the Hurwitz action contains an element of a certain explicitly specified finite set. The purpose of the present paper is to give an easy criterion to decide if two given elements of this finite set belong to the same orbit. To give precise statements, we need to introduce some notation with slightly differs from that in [4]. Let us extend the alphabet A up to = A∪{σ 0 ,σ 1 ,σ 2 }. Let A * and * be the free monoids generated by A and respectively. If U, V ∈ * , then U ≡ V stands for equality in * (i. e. letterwise coincidence of words) and U = V (when U, V ∈ A * ) stands for equality of the corresponding elements of B 3 . For U ∈ * , we denote the word obtained from U by erasing of all lettersσ i (resp. by replacing eachσ i with σ i or by replacing each σ i withσ i ) byŪ (resp. U ′ orÛ ).It is easy to check that any X ∈ B 3 can be written asIf, moreover, U does not contain any subword which is equal (in B 3 ) to δ, then the presentation of X in the form (1) is unique and it is called the right Garside normal form.If W ′ = δ p , then [W ] stands for the quasipositive factorization ofW δ −p of the form (X 1 , . . . , X k ) where X i = A i x i A −1 i , A i = W 1 . . . W i . To each X ∈ B 3 we associate a graph G 0 (X) in the following way. Let (1) be the right Garside normal form of X. We define the set of vertices of G 0 (X) as