1: \begin{abstract}

2: A spread-out lattice animal is a finite connected set of edges in 

3: $\{\{x,y\}\subset\Zd:0<\|x-y\|\le L\}$.  A lattice tree is a lattice 

4: animal with no loops.  Let $\pc$ be the model-dependent radius of convergence 

5: of (the sum of) the 2-point function $\tau_p(x)$, which is a generating 

6: function of lattice trees or animals containing the origin $o$ and 

7: $x\in\Zd$, with fugacity $p/|\Lambda|$, where $|\Lambda|$ is the degree of a 

8: vertex.  The best estimate on $\pc$ so far was achieved by Penrose \cite{p94}: 

9: $\pc=1/e+O(L^{-2d/7}\log L)$ for both models for all $d\ge1$.  In this paper, 

10: we show that $\pc=1/e+CL^{-d}+O(L^{-d-1})$ for all $d>8$, where the 

11: non-universal constant $C$ has the random-walk representation

12: \begin{align*}

13: C_\mathrm{LT}=\sum_{n=2}^\infty\frac{n+1}{2e}U^{*n}(o),&&

14: C_\mathrm{LA}=C_\mathrm{LT}-\frac1{2e^2}\sum_{n=3}^\infty U^{*n}(o),

15: \end{align*}

16: %\begin{align*}

17: %C=\frac3{2e}U^{*2}(o)+

18: % \begin{cases}

19: % \dpst\sum_{n=3}^\infty\frac{n+1}{2e}U^{*n}(o)&[\text{lattice trees}],\\[15pt]

20: % \dpst\sum_{n=3}^\infty\bigg(\frac{n+1}{2e}-\frac1{2e^2}\bigg)U^{*n}(o)

21: %  &[\text{lattice animals}],

22: % \end{cases}

23: %\end{align*}

24: where $U^{*n}$ is the $n$-fold convolution of the uniform distribution on the 

25: $d$-dimensional ball $\{x\in\R^d:\|x\|\le1\}$.  The proof is based on a novel 

26: use of the lace expansion for $\tau_p(x)$ and detailed analysis of the 1-point 

27: function at a certain value of $p$ that is designed to make the analysis 

28: extremely simple.

29: %\blue{The point for our proof is that we prove $p_c-p_1$ is order $L^{-2d}$, 

30: %where $p_1$ is defined for the both models by the identity $g_{p_1}p_1=1$. 

31: %Since $p_c-p_1$ is irrelevant for the first error term which is order 

32: %$L^{-d}$, it suffices to get the coefficient that we investigate $p_1$. 

33: %Due to the identity $g_{p_1}p_1=1$, in which the lace expansion coefficients 

34: %doesn't appear, it enables overwhelmingly easy estimate of $p_c$.}

35: \end{abstract}

36: