A loop of order $n$ possesses at least $3n^2-3n+1$ associative triples. However, no loop of order $n>1$ that achieves this bound seems to be known. If the loop is involutory, then it possesses at least $3n^2-2n$ associative triples. Involutory loops with $3n^2-2n$ associative triples can be obtained by prolongation of certain maximally nonassociative quasigroups whenever $n-1$ is a prime greater than or equal to $13$ or $n-1=p^{2k}$, $p$ an odd prime. For orders $n\le 9$ the minimum number of associative triples is reported for both general...

Let $G$ be a finite group with a dicyclic subgroup $H$. We show that if there exist $H$-connected transversals in $G$, then $G$ is a solvable group. We apply this result to loop theory and show that if the inner mapping group $I\left(Q\right)$ of a finite loop $Q$ is dicyclic, then $Q$ is a solvable loop. We also discuss a more general solvability criterion in the case where $I\left(Q\right)$ is a certain type of a direct product.

This paper completely solves the isomorphism problem for Moufang loops $Q=GC$ where $G⊴Q$ is a noncommutative group with cyclic subgroup of index two and $\left|Z\right(G\left)\right|\le 2$, $C$ is cyclic, $G\cap C=1$, and $Q$ is finite of order coprime to three.

A binary operation “$·$” which satisfies the identities $x·e=x$, $x·x=e$, $(x·y)·x=y$ and $x·y=y·x$ is called a Steiner loop. This paper revisits the proof of the necessary and sufficient conditions for the existence of a Steiner loop of order $n$ with centre of order $m$ and discusses the connection of this problem to the question of the maximum number of Pasch configurations which can occur in a Steiner triple system (STS) of a given order. An STS which attains this maximum for a given order is said to be . We show that...

For G = SU(n), Sp(n) or Spin(n), let $C_G\left(SU\left(2\right)\right)$ be the centralizer of a certain SU(2) in G. We have a natural map $J:G/C_G\left(SU\left(2\right)\right)\to \Omega ₀³G$. For a generator α of $H⁎(G/C_G\left(SU\left(2\right)\right);\mathbb{Z}/2)$, we describe J⁎(α). In particular, it is proved that $J⁎:H⁎(G/C_G\left(SU\left(2\right)\right);\mathbb{Z}/2)\to H⁎(\Omega ₀³G;\mathbb{Z}/2)$ is injective.

We study identities of the form $$L_{x_0}{\varphi }_1\cdots {\varphi }_n{R}_{x_{n+1}}=R_{x_{n+1}}{\varphi }_{\sigma \left(1\right)}\cdots {\varphi }_{\sigma \left(n\right)}{L}_{x_0}$$ in quasigroups, where $n\ge 1$, $\sigma$ is a permutation of $\{1,...,n\}$, and for each $i$, ${\varphi }_i$ is either $L_{x_i}$ or $R_{x_i}$. We prove that in a quasigroup, every such identity implies commutativity. Moreover, if $\sigma$ is chosen randomly and uniformly, it also satisfies associativity with probability approaching $1$ as $n\to \infty$.

For a finite group $G$ denote by $N\left(G\right)$ the set of conjugacy class sizes of $G$. In 1980s, J. G. Thompson posed the following conjecture: If $L$ is a finite nonabelian simple group, $G$ is a finite group with trivial center and $N\left(G\right)=N\left(L\right)$, then $G\cong L$. We prove this conjecture for an infinite class of simple groups. Let $p$ be an odd prime. We show that every finite group $G$ with the property $Z\left(G\right)=1$ and $N\left(G\right)=N\left(A_i\right)$ is necessarily isomorphic to $A_i$, where $i\in \{2p,2p+1\}$.

Let $f$, $g$ and $h$ be three distinct primitive holomorphic cusp forms of even integral weights $k_1$, $k_2$ and $k_3$ for the full modular group $\Gamma =\mathrm{SL}(2,\mathbb{Z})$, respectively, and let ${\lambda }_f\left(n\right)$, ${\lambda }_g\left(n\right)$ and ${\lambda }_h\left(n\right)$ denote the $n$th normalized Fourier coefficients of $f$, $g$ and $h$, respectively. We consider the cancellations of sums related to arithmetic functions ${\lambda }_g\left(n\right)$, ${\lambda }_h\left(n\right)$ twisted by ${\lambda }_f\left(n\right)$ and establish the following results: $$\sum _{n\le x}{\lambda }_f\left(n\right){\lambda }_g{\left(n\right)}^i{\lambda }_h{\left(n\right)}^j{\ll }_{f,g,h,\epsilon }{x}^{1-1/2^{i+j}+\epsilon }$$ for any $\epsilon >0$, where $1\le i\le 2$, $j\ge 5$ are any fixed positive integers.

It is well known that given a Steiner triple system one can define a quasigroup operation $·$ upon its base set by assigning $x·x=x$ for all $x$ and $x·y=z$, where $z$ is the third point in the block containing the pair $\{x,y\}$. The same can be done for Mendelsohn triple systems, where $(x,y)$ is considered to be ordered. But this is not necessarily the case for directed triple systems. However there do exist directed triple systems, which induce a quasigroup under this operation and these are called Latin directed...