Gyrogroups and homogeneous loops (Q1970535)

The author gives firstly a survey on connections between the gyrogroups of A. A. Ungar, the Bruck loops and \(K\)-loops. Then he considers properties which Ungar established by studying the relativistic velocity addition. Let \((L,+)\) be a groupoid with a neutral element \(0\) such that \(\forall a\in L \exists_1-a\in L\): \(a+(-a)=(-a)+a=0\). For \(a\in L\) let \(a^+\colon L\to L\); \(x\mapsto a+x\). Then, following Ungar, the author calls \((L,+)\) a gyrogroup if for all \(a,b\in L\) the following properties are satisfied: (1) There is a map \(\delta_{a,b}\colon L\to L\) with \(a^+\circ b^+=(a+b)^+\circ\delta_{a,b}\), (2) \((a+b)^+=(a+b)^+\circ\delta_{a,b}\circ\delta_{b,a}\), (3) \(\delta_{0,a}=\text{id}\), (4) \(\delta_{a+b,b}=\delta_{a,b}=\delta_{a,b+a}\). For a gyrogroup the author proves \(-(u+v)=\delta_{u,v}(-v+(-u))\) (Theorem 2) and claims: If ``(5) \(a+b=\delta_{a,b}(b+a)\)'' is satisfied then \(-(u+v)=-u+(-v)\) (Lemma 4) which I could not verify! If (5) is valid then \((x+y)^+=\delta_{x,y}\circ y^+\circ x^+\) (Theorem 3) which I could not verify: the author uses further operations \((x,y)\mapsto xy\) and \(x\mapsto x^2\) which he does not define. Then he proves: Let \((L,+)\) be a loop (then \(\delta_{a,b}=((a+b)^+)^{-1}\circ a^+\circ b^+)\) such that ``(6) \(\forall a,b\in L\); \(\delta_{a,b}\) is an automorphism of \((L,+)\)'' and ``(7) \((a^+)^{-1}=(-a)^+\)'' then: (a) If ``(8) \(\forall a,b\in L\), \(-(a+b)=-a+(-b)\)'' is valid then also (2) (Theorem 4), (b) If (8) and ``(9) \(\forall a,b\in L\): \(\delta_{b,a+b}\circ\delta_{a,b}=\text{id}\)'' are valid (then \((L,+)\) is a Bruck loop since (9) describes the Bol loops!) then \((L,+)\) is a gyrogroup (Theorem 5). (c) If ``(4L) \(\forall a,b\in L\): \(\delta_{a,b}=\delta_{a+b,b}\)'' is valid then the identities: \(\delta_{x,y}=\delta_{-x,-y+(-x)}=\delta_{-x,-y}=(\delta_{y,x})^{-1}\), \(\delta_{x+y,y}=\delta_{y,-y+(-x)}\), \(\delta_{-x,x+y}=\delta_{y,x+y}=\delta_{y,x}\) are satisfied. If (7) and (4L) are valid then \((L,+)\) is a Bol loop and if \(f\in\Aut(L,+)\) then \(f\circ\delta_{x,y}\circ f^{-1}=\delta_{f(x),f(y)}\). Finally, the author states: A gyrogroup \((L,+)\) is a loop with (6), (7) and (4L) (Theorem 8). In the proof I am missing: why is \((L,+)\) a loop?, why \(\delta_{x,y}\in\Aut(L,+)\)? where is (7) deduced from (1), (2), (3), (4)?