\documentclass[twoside]{article} \usepackage{amssymb} % used for R in Real numbers \pagestyle{myheadings} \markboth{\hfil Multiplicity of solutions for boundary-value problems\hfil EJDE--1999/21} {EJDE--1999/21\hfil Idris Addou \hfil} \begin{document} \title{\vspace{-1in}\parbox{\linewidth}{\footnotesize\noindent {\sc Electronic Journal of Differential Equations}, Vol. {\bf 1999}(1999), No.~21, pp. 1--27. \newline ISSN: 1072-6691. URL: http://ejde.math.swt.edu or http://ejde.math.unt.edu \newline ftp ejde.math.swt.edu \quad ftp ejde.math.unt.edu (login: ftp)} \vspace{\bigskipamount} \\ % Multiplicity of solutions for quasilinear elliptic boundary-value problems \thanks{ {\em 1991 Mathematics Subject Classifications:} 34B15. \hfil\break\indent {\em Key words and phrases:} Multiple solutions, elliptic boundary-value problems. \hfil\break\indent \copyright 1999 Southwest Texas State University and University of North Texas. \hfil\break\indent Submitted January 21, 1999. Published June 16, 1999.} } \date{} % \author{Idris Addou } \maketitle \begin{abstract} This paper is concerned with the existence of multiple solutions to the boundary-value problem $$-(\varphi_p(u') ) '=\lambda \varphi_q(u) +f(u)\quad\mbox{in } (0,1)\,,\quad u(0) =u(1) =0\,,$$ where $p,q>1$, $\varphi_x(y) =|y|^{x-2}y$, $\lambda $ is a real parameter, and $f$ is a function which may be sublinear, superlinear, or asymmetric. We use the time map method for showing the existence of solutions. \end{abstract} \newtheorem{theorem}{Theorem} \newtheorem{proposition}[theorem]{Proposition} \newtheorem{lemma}[theorem]{Lemma} \section{Introduction} \label{sec1} We study the existence and multiplicity of solutions for the boundary-value problem \begin{eqnarray} \label{p1} &-(\varphi_p(u') ) '=\lambda \varphi_q(u) +f(u) \quad\mbox{in }(0,1)&\\ &u(0) =u(1) =0\,,& \nonumber \end{eqnarray} where $p,q>1$, $\varphi_x(y) =| y| ^{x-2}y$, $\lambda \in {\mathbb R}$, and $f$ is a continuous function such that $$ a_\pm =\lim_{s\to \pm \infty }f(s) /\varphi_q(s) \quad\mbox{and}\quad a_0=\lim_{s\to 0}f(s)/\varphi_q(s) $$ exist as real numbers. Also, we assume that $m_\pm :=\inf_{\pm s\geq 0}f(s) /\varphi_q(s)$ exists in ${\mathbb R}$, and define $$ m:=\inf_{u\in {\mathbb R}}f(s) /\varphi_q(s) =\min (m_{+},m_{-}) \quad\mbox{and}\quad m_k^{\pm}= \cases{ m_\pm & if $k=1$ \cr m & if $k\geq 2$.} $$ We shall consider the following three cases: Superlinear case, $a_0<\min (a_{-},a_{+})$; Sublinear case, $a_0>\max (a_{-},a_{+})$; and Asymmetric case, $a_{-}1$, we mention a recent paper by J. Wang \cite{W} where positive solutions are studied. Now we define some sets that will be used in the statement of the main results. For $k\geq 1$, let $$ S_k^+=\left\{ \begin{array}{c} u\in C^1([0,1]) :u \mbox{ admits exactly }( k-1) \mbox{ zeros in }(0,1) , \\ \mbox{all are simple, }u(0) =u(1) =0\mbox{ and }u'(0) >0 \end{array} \right\}\,, $$ then $S_k^-=-S_k^+$ and $S_k=S_k^+\cup S_k^-$. If $u\in C([0,1]) $ is a real-valued function vanishing at $x_1$ and $x_2$ and not between them, (with $x_11$, $\lambda_k=k^p\lambda_1$ and $$ \lambda_1=(p-1) (2\int_0^1(1-t^p) ^{-1/p}dt) ^p=(p-1) (\frac{2\pi }{p\sin (\pi /p) }) ^p. $$ For fixed real constants $a_{-}$ and $a_{+}$, consider the boundary value problem \begin{eqnarray}\label{p3} &-(\varphi_p(u') ) ' = \lambda \varphi_p(u) +a_{+}\varphi_p(u^+) -a_{-}\varphi _p(u^-) ,\mbox{in }(0,1) & \\ &u(0) = u(1) =0\,,& \nonumber \end{eqnarray} where $\lambda $ is a real parameter. If $\lambda $ is such that problem (\ref{p3}) admits a nontrivial solution $u_\lambda $, then $\lambda $ is called a half-eigenvalue of (\ref{p3}). In the particular case where $p=2$, this definition goes back to Berestycki \cite{Berestycki}. For any integer $k\geq 1$, let $$ b_k^\pm =\left\{ \begin{array}{ll} a_\pm & \mbox{if }k=1 \\ \min (a_{-},a_{+}) & \mbox{if }k\geq 2 \end{array} \right. \quad\mbox{and}\quad c_k^\pm =\left\{ \begin{array}{ll} a_\pm & \mbox{if }k=1 \\ \max (a_{-},a_{+}) & \mbox{if }k\geq 2\,. \end{array} \right. $$ \begin{proposition} \label{prop1} For fixed real constants $a_{-}$ and $a_{+}$, the set of half-eigenvalues of problem (\ref{p3}) consists of two increasing sequences $(\lambda_k^+)_{k\geq 1}$ and $(\lambda_k^-) _{k\geq 1}$, satisfying $h_k^\pm (\lambda_k^\pm ) =1$, for all $k\geq 1$, where $$\displaylines{ h_{2n}^\pm (\lambda ) :=\frac{\lambda_n^{1/p}}{(a_{\pm }+\lambda ) ^{1/p}}+\frac{\lambda_n^{1/p}}{(a_{\mp }+\lambda ) ^{1/p}},\mbox{ for all }\lambda >-b_{2n}^\pm ,\ n\geq 1, \cr h_{2n+1}^\pm (\lambda ) :=\frac{\lambda_{n+1}^{1/p}}{( a_\pm +\lambda ) ^{1/p}}+\frac{\lambda_n^{1/p}}{(a_{\mp }+\lambda ) ^{1/p}},\mbox{ for all }\lambda >-b_{2n+1}^\pm ,\ n\geq0\,, } $$ with the convention $\lambda_0=0$. Moreover, if $a_{\mp }c_k^\pm $ and $\max (-m_k^\pm ,\lambda _k-a_0) <\lambda_k^\pm $, problem (\ref{p1}) admits at least a solution in $A_k^\pm $ for all $\lambda $ satisfying $\max (-m_k^\pm ,\lambda_k-a_0) <\lambda <\lambda_k^\pm$. \end{enumerate} \end{theorem} The next result deals with the asymmetric case. For any integer $n\geq 1$, let $$ a_{2n+1}^\pm =\frac{na_{\mp }+(n+1) a_\pm }{2n+1}\quad \mbox{and}\quad a_{2n}^\pm =\frac{na_{-}+na_{+}}{2n}=\frac{a_{-}+a_{+}}2\,. $$ \begin{theorem} \label{thm2} Assume that $q=p$ and $a_{-}-m_k^\pm $. In particular, (\ref{p1}) admits infinitely many solutions for all $\lambda >-m$. \end{theorem} The paper is organized as follows. The next section is dedicated to the proof of Proposition \ref{prop1}. In Section \ref{sec3}, we present the method used for proving the main results of this paper. In Section \ref{sec4}, we present some preliminary lemmas. In Section \ref{sec5}, we prove the main results, present some remarks. The paper ends with an appendix which contains a brief historical overview on time maps. \section{Proof of Proposition \protect{\ref{prop1}}} \label{sec2} Notice that for all $u\in {\mathbb R},\varphi _p(u) =\varphi_p(u^+) -\varphi_p(u^-) $; so, problem (\ref{p3}) may be written as \begin{eqnarray} \label{p4} &-(\varphi_p(u') ) ' = \alpha _{+}\varphi_p(u^+) -\alpha_{-}\varphi_p(u^-),\quad \mbox{in }(0,1)\,,& \\ &u(0) = u(1) =0\,,& \nonumber \end{eqnarray} with $\alpha_\pm =\lambda +a_\pm $. The set of $(\alpha _{+},\alpha_{-}) \in {\mathbb R}^2$ such that problem (\ref {p4}) has a nontrivial solution is known as the Fu\v cik spectrum for the operator $(\varphi_p(u') ) '$ under Dirichlet boundary conditions. See for instance, Dr\'abek \cite {Drabek1}, Boccardo et al. \cite{Boccardo}, Huang and Metzen \cite{Huang}. We denote this spectrum by $\sum_p$. One has $\sum_p=\bigcup_{k\geq 0}C_k^\pm $, where \begin{eqnarray*} &C_0^\pm =\left\{ (\alpha_{+},\alpha_{-}) \in {\mathbb R}^2:\alpha_\pm =\lambda_1\right\} ,& \\ &C_{2n}^\pm =\left\{ (\alpha_{+},\alpha_{-}) \in ( {\mathbb R}^+) ^2:(\lambda_n^{1/p}/\alpha_{\pm }^{1/p}) +(\lambda_n^{1/p}/\alpha_{\mp }^{1/p}) =1\right\} ,n\geq 1,& \\ &C_{2n+1}^\pm =\left\{ (\alpha_{+},\alpha_{-}) \in ( {\mathbb R}^+) ^2:(\lambda_{n+1}^{1/p}/\alpha_{\pm }^{1/p}) +(\lambda_n^{1/p}/\alpha_{\mp }^{1/p}) =1\right\} ,n\geq 0\,.& \end{eqnarray*} So, $\lambda $ is a half-eigenvalue of (\ref{p3}) if and only if $( \lambda +a_{+},\lambda +a_{-}) \in \sum_p$, that is, $h_k^{\pm }(\lambda ) =1,$ for some $k\geq 1$. Since the function $\lambda \mapsto h_k^\pm (\lambda ) ,k\geq 1$, is strictly decreasing on $(-b_k^\pm ,+\infty ) $ and its limits at $-b_k^\pm $ and $+\infty $ are $+\infty $ and $0$ respectively, it follows that the equation $h_k^\pm (\lambda ) =1$ admits a unique solution $\lambda =\lambda_k^\pm $. It is easy to check that for $a_{-}-a_{-}$ one has $$ h_{2n-1}^+(\lambda ) -a_{-}$, $k\geq 1$, $$ \frac{\lambda_k^{1/p}}{(a_{+}+\lambda ) ^{1/p}}<\frac{\lambda _k^{1/p}}{(a_{-}+\lambda ) ^{1/p}},\quad\mbox{so}\quad\frac{\lambda _{2k}^{1/p}}{(a_{+}+\lambda ) ^{1/p}}0$ and $\kappa =+,-$, let $$ X_\kappa (E) =\left\{ s\in {\mathbb R}:\kappa s\geq 0\quad\mbox{and}\quad E^p-p'G(\xi ) >0,\,\forall \,\xi ,0\leq \kappa \xi <\kappa s\right\} $$ and $$ r_\kappa (E) =\cases{ 0 & if $X_\kappa (E) =\emptyset$ \cr \kappa \sup (\kappa X_\kappa (E) ) & otherwise. } $$ Note that $r_\kappa $ may be infinite. We shall also make use of the following sets, $$ D_\kappa =\left\{ E>0:0<| r_\kappa (E) | <+\infty \mbox{ and }\kappa g(r_\kappa (E) ) >0\right\}$$ and $D=D_{+}\cap D_{-}$. Also, let $D_k^\kappa :=D$ if $k\geq 2$, and $D_1^\kappa :=D_\kappa $. Define the following time-maps, \begin{eqnarray*} &T_\kappa (E) = \kappa \int_0^{r_\kappa (E) }(E^p-p'G(t) ) ^{-1/p}dt, \quad E\in D_\kappa \,,&\\ &T_{2n}^\kappa (E) = n(T_{+}(E) +T_{-}(E) ) , \quad n\in {\mathbb N},\ E\in D\,,&\\ &T_{2n+1}^\kappa (E) = T_{2n}^\kappa (E) +T_\kappa (E) , \quad n\in {\mathbb N}, \ E\in D. \end{eqnarray*} \begin{theorem}[Quadrature method] \label{quad} Assume that (\ref{meth2}) holds. Let $E>0$, $k\in {\mathbb N}^{*}$, $\kappa =+,-$. If $E\in D_k^\kappa $ and $T_k^\kappa ( E) =1/2$, problem (\ref{meth3}) admits at least a solution $u_k^\kappa \in A_k^\kappa $ satisfying $(u_k^\kappa ) '( 0) =\kappa E$, and this solution is unique. \end{theorem} This theorem is well-known, but we did not find a convenient reference to the precise statement used later. The paper by Guedda and Veron \cite{GueddaVeron}, seems to be the first one dealing with time maps approach when the differential operator is the one dimensional p-Laplacian. An easy adaptation of the ideas contained in \cite{GueddaVeron} and in the paper by Del Pino and Manasevich \cite{DelpinoMana1}, allows one to prove Theorem \ref{quad}. Also, we mention the papers by Manasevich and Zanolin \cite{ManaZanolin} and Manasevich et al. \cite{ManasevichetAl}, where one can find time maps used when the differential operator is the one dimensional p-Laplacian. Notice that time maps were also used when the differential operator generalizes the p-Laplacian; see for example Garcia-Huidobro et al. \cite{GarciaetAl1}, \cite{GarciaetAl2}, \cite {GarciaetAl3}, \cite{GarciaetAl4}, \cite{GarciaetAl5}, Garcia-Huidobro and Ubilla \cite{GarciaUbilla}, Huang and Metzen \cite{Huang} and Ubilla \cite {Ubilla}. A brief historical overview of time maps is given in the appendix. For the sake of completeness, we dedicate the rest of this section to the proof of Theorem \ref{quad} for the case where $k=1$ and $\kappa =+$. The adaptation of the other cases may be handled similarly. We shall say that $u$ is a solution of (\ref{meth3}) if $u$ and $\varphi _p(u') $ belong to $C^1([0,1]) $ and $u$ satisfies $$ -(\varphi_p(u'(x) ) ) ^{\prime }=g(u(x) ) \ \forall x\in (0,1)\,,\quad\mbox{with}\quad u(0) =u(1) =0\,. $$ Let us assign to each function $u\in C^1([0,1]) $ the set $$ Z(u') =\left\{ x\in [0,1]:u^{\prime }(x) =0\right\} . $$ \begin{lemma} \label{Lem1}Assume that (\ref{meth2}) holds and $u$ is a solution of (\ref {meth3}). Then $u\in C^2([0,1]) $ if $12$. \end{lemma} \paragraph{Proof.} The identity $t=\varphi_{p'}o\varphi _p(t) $ for all $t\in {\mathbb R}$ implies that $$ u'(x) =\varphi_{p'}o(\varphi_p( u') ) (x) \mbox{\ \ for\ all\ }x\in [0,1]. $$ Thus, if $12$, it follows that $11$ and assume that $u$ is a solution of problem (\ref {meth3}). Then $$ (| u'| ^p(x) +p'G( u(x) ) ) '=0,\mbox{\ \ \ for all }x\in [0,1]. $$ \end{lemma} \paragraph{Proof.} Let $x_0\in Z(u') $. One has $$ (G(u))' (x_0) =g(u(x_0) ) u'(x_0) =0\,. $$ Let us prove that \begin{equation} (| u'| ^p) '(x_0) =0\,. \label{Am} \end{equation} One has for all $x\neq x_0$, $$ \frac{| u'(x) | ^p-| u'( x_0) | ^p}{x-x_0}=u'(x) \times \frac{\varphi_p(u'(x) ) -\varphi_p( u'(x_0) ) }{x-x_0}. $$ Thus, $\lim_{x\to x_0}u'(x) =u'(x_0) =0$, and $$ \lim_{x\to x_0}\frac{\varphi_p(u'(x) ) -\varphi_p(u'(x_0) ) }{x-x_0}=(\varphi_p(u') ) '( x_0) =-g(u(x_0) ) \in {\mathbb R}. $$ Therefore, (\ref{Am}) is proved. Regarding (\ref{AD}), Lemma \ref{Lem2} follows. \hfill$\diamondsuit$ \paragraph{Remark.} In Lemmas \ref{Lem1} and \ref{Lem2}, the solutions, $u$, are arbitrary and are not necessarily in $A_1^+$. \smallskip Now, assume that $u$ is a solution of problem (\ref{meth3}) belonging to $A_1^+$. Thus, \begin{equation} u'>0\mbox{\ in\ }[0,(1/2) ) \mbox{\ \ and\ \ }u'(1/2) . \label{kyd} \end{equation} It follows that $$ \sup \left\{ x\in [0,1) :u'(t) >0,\forall t\in [0,x) \right\} =1/2, $$ and by the energy relation one gets \begin{equation} u'(x) =\left\{ (u'(0) ) ^p-p'G(u(x) ) \right\} ^{1/p}\mbox{,\ \ for all }x\in [0,1]. \label{jet} \end{equation} Thus, $$ \sup \left\{ x\in [0,1) :(u'(0) ) ^p-p'G(u(t) ) >0,\forall t\in [0,x) \right\} =1/2, $$ or equivalently $$ \sup \left\{ x\in [0,1) :(u'(0) ) ^p-p'G(z) >0,\forall z\in [0,u( x) ) \right\} =1/2, $$ which implies that $$ \sup \left\{ s\geq 0:(u'(0) ) ^p-p^{\prime }G(\xi ) >0,\forall \xi \in [0,s) \right\} =u(1/2) . $$ Also, by (\ref{jet}) it follows that \begin{equation} x=\int_0^{u(x) }\left\{ (u'(0) ) ^p-p'G(\xi ) \right\} ^{-1/p}d\xi \mbox{,\ \ \ for all }x\in [0,(1/2) ]. \label{qsfj} \end{equation} Thus, the improper integral in (\ref{qsfj}) is convergent for all $x\in [0,(1/2) ]$ and in particular, $u'( 0) $ is such that the improper integral $$ \int_0^{r_{+}(u'(0) ) }\left\{ ( u'(0) ) ^p-p'G(\xi ) \right\} ^{-1/p}d\xi $$ converges and is equal to $(1/2) $, where for all $E>0 $$$ r_{+}(E) =\sup X_{+}(E) \mbox{\ if }X_{+}( E) \neq \emptyset \mbox{\ \ and\ \ }r_{+}(E) =0\mbox{\ if }X_{+}(E) =\emptyset $$ and $$ X_{+}(E) =\left\{ s\geq 0:E^p-p'G(\xi ) >0,\forall \xi \in [0,s) \right\} . $$ It follows that if $u$ is a solution of problem (\ref{meth3}) belonging to $A_1^+$, then there exists $E_{*}\in \tilde D_{+}$ with $$ \tilde D_{+}=\big\{ E>0:00\mbox{ for all }u\in (0,r_{+}( E_{*}) ) . $$ Let $u_{+}$ be the inverse of $h_{+}$ defined by $$ u_{+}(x) =h_{+}^{-1}(x) \in [0,r_{+}( E_{*}) ],\mbox{ for all }x\in [0,(1/2)]\,, $$ and let $u$ be defined on $[0,1]$ by $$ u(x) =\left\{ \begin{array}{l} u_{+}(x) \mbox{\ if\ }x\in [0,(1/2) ] \\ u_{+}(1-x) \mbox{\ if }x\in [(1/2) ,1] . \end{array} \right. $$ It easy to show that this function $u$ is a solution of problem (\ref{meth3}) belonging to $A_1^+$ and satisfying $u'(0) =E_{*}$, $\max_{[0,1]}u=u(1/2) =r_{+}(E_{*}) $. Let us prove its uniqueness. Assume that $v$ is also a solution of problem (\ref{meth3}) belonging to $A_1^+$ and satisfies $$ v'(0) =E_{*},\max_{[0,1]}v=v(1/2) =v_{+}(E_{*}) . $$ By (\ref{qsfj}) it follows that $$ x=\int_0^{u(x) }\left\{ E_{*}^p-p'G(\xi ) \right\} ^{-1/p}d\xi =\int_0^{v(x) }\left\{ E_{*}^p-p^{\prime }G(\xi ) \right\} ^{-1/p}d\xi , $$ for all $x\in [0,1/2]$. Thus, $$ \int_{u(x) }^{v(x) }\left\{ E_{*}^p-p^{\prime }G(\xi ) \right\} ^{-1/p}d\xi =0,\mbox{\ \ for all }x\in [ 0,1/2]. $$ Thus, $u=v$ on $[0,1/2]$, and by symmetry it follows that $u=v$ on $[0,1]$. Therefore, because $D_{+}\subset \tilde D_{+}$, Theorem \ref{quad} is proved for the case $k=1$ and $\kappa =+$. \hfill$\diamondsuit$\smallskip \section{Preliminary Lemmas} \label{sec4}In order to define the time maps we need the following. \begin{lemma} \label{lemma1} For $s\in {\mathbb R}$, consider the equation \begin{equation} E^p-p'(F(s) +\lambda \Phi_q(s) ) =0, \label{Eq.1} \end{equation} where $p>1$, $E\geq 0$ and $\lambda \in {\mathbb R}$ are real parameters, $F(s) =\int_0^sf(t) dt$ and $\Phi _q(s) =\int_0^s\varphi_q(t) dt$. If $\lambda >-m_\pm $, then for any $E>0$, equation (\ref{Eq.1}) admits a unique positive zero $s_{+}=s_{+}(\lambda ,E) $ (resp. a unique negative zero $s_{-}=s_{-}(\lambda ,E) $) and if $E=0$ it admits no positive (resp. negative) zero beside the trivial one $s_\pm =0$. Moreover, for all $\lambda >-m_\pm ,p>1$, \begin{description} \item[(i)] The function $E\longmapsto s_\pm (\lambda ,E) $ is $C^1$ on $(0,+\infty ) $, and $$ \pm \frac{\partial s_\pm }{\partial E}(\lambda ,E) =\frac{\pm (p-1) E^{p-1}}{f(s_\pm (\lambda ,E) ) +\lambda \varphi_q(s_\pm (\lambda ,E) ) }>0,\ \forall E>0\,. $$ \item[(ii)] $\lim\limits_{E\to 0^+}s_\pm (\lambda ,E) =0$, $\lim\limits_{E\to +\infty }s_\pm ( \lambda ,E) =\pm \infty $. \item[(iii)] $\lim\limits_{E\to 0}\frac{E^p}{| s_\pm ( \lambda ,E) | ^q}=\frac{p'}q(a_0+\lambda )$, $\lim\limits_{E\to +\infty }\frac{E^p}{| s_{\pm }(\lambda ,E) | ^q}=\frac{p'}q(a_{\pm }+\lambda ) $. \item[(iv)] $\lim\limits_{E\to 0}\frac E{| s_{\pm }(\lambda ,E) | }=\cases{ 0 & if $q-p>0$ \cr ((a_0+\lambda ) /(p-1) ) ^{1/p} & if $ q-p=0$ \cr +\infty & if $ q-p<0$, }$ $\lim\limits_{E\to +\infty }\frac E{| s_\pm (\lambda ,E) | }=\cases{ +\infty & if $ q-p>0$\cr ((a_\pm +\lambda ) /(p-1) ) ^{1/p} & if $q-p=0$\cr 0 & if $q-p<0$.} $ \item[(v)] $\lim\limits_{E\to 0}\frac{F(s_{\pm }(\lambda ,E) t) }{E^p}=\frac{t^p}{p'}\frac{a_0}{a_0+\lambda },\forall t>0$, $\lim\limits_{E\to +\infty }\frac{F(s_\pm (\lambda ,E) t) }{E^p}=\frac{t^p}{p'}\frac{a_\pm }{a_{\pm }+\lambda },\forall t>0$. \end{description} \end{lemma} \paragraph{Proof.} For $p>1$ and $E\geq0$ fixed, consider the function $$ s\longmapsto G_\pm (\lambda ,E,s) :=E^p-p'( F(s) +\lambda \Phi_q(s) ) , $$ defined in ${\mathbb R}^\pm $ and strictly decreasing on $(0,+\infty ) $ (resp. strictly increasing on $(-\infty,0) $), because $$ \frac{dG_\pm }{ds}(\lambda ,E,s) =-p'\varphi _q(s) (\frac{f(s) }{\varphi_q(s) }+\lambda ) \quad\mbox{and}\quad m_\pm +\lambda >0\,. $$ One has $G_\pm (\lambda ,E,0) =E^p\geq 0$, and via l'Hospital's rule, \begin{eqnarray*} \lim\limits_{s\to +\infty }G_\pm (\lambda ,E,s) & = & \lim\limits_{s\to +\infty }E^p-p'\Phi_q(s) (\frac{F(s) }{\Phi_q(s) }+\lambda ) \\ & = & E^p-p'\lim\limits_{s\to +\infty }\Phi_q( s) (\lim\limits_{s\to +\infty }\frac{f(s) }{\varphi_q(s) }+\lambda ) \\ & = & -\infty . \end{eqnarray*} So, it is clear that in the case $m_{+}+\lambda >0$ (resp. $m_{-}+\lambda >0$) for any $E>0$, (\ref{Eq.1}) admits a unique positive zero, $s_{+}=s_{+}(\lambda ,E) $, (resp. a unique negative zero, $s_{-}=s_{-}(\lambda ,E) $), and if $E=0$, it admits no positive (resp. negative) zero beside the trivial one $s=0$. \smallskip Now, for any $p>1$ and $\lambda >-m_\pm $, consider the real valued function, $$ (E,s) \longmapsto G_\pm (E,s) :=E^p-p^{\prime }(F(s) +\lambda \Phi_q(s) ) , $$ defined on $\Omega_{+}=(0,+\infty ) ^2$ (resp. $\Omega _{-}=(0,+\infty ) \times (-\infty ,0) $). One has $G_\pm \in C^1(\Omega_\pm ) $ and, $$ \frac{\partial G_\pm }{\partial s}(E,s) =-p'\varphi _q(s) (\frac{f(s) }{\varphi_q(s) }+\lambda ) \quad\mbox{in }\Omega_\pm \,; $$ hence, because $m_\pm +\lambda >0$, it follows that $$ \pm \frac{\partial G_\pm }{\partial s}(E,s) <0,\mbox{\ \ in \ }\Omega_\pm , $$ and one may observe that $s_\pm (\lambda ,E) $ belongs to the open interval $(0,+\infty ) $ (resp. $(-\infty ,0) $) and from its definition satisfies \begin{equation} G_\pm (E,s_\pm (\lambda ,E) ) =0\,. \label{Eq.2} \end{equation} So, one can make use of the implicit function theorem to show that the function $E\mapsto s_\pm (\lambda ,E) $ is $C^1(( 0,+\infty ) ,{\mathbb R}) $ and to obtain the expression of $\frac{\partial s_\pm }{\partial E}(\lambda ,E) $ given in {\bf (i)}. Hence, by $m_\pm +\lambda >0$, it follows that for any fixed $p>1$ and $\lambda >-m_\pm $, the function defined on $(0,+\infty ) $ by $E\mapsto s_\pm (\lambda ,E) $ is strictly increasing (resp. strictly decreasing) and bounded from below by $0$ (resp. by $-\infty $) and from above by $+\infty $ (resp. by $0$). Then the limit $\lim_{E\to 0^+}s_\pm (\lambda ,E) =\ell_0^\pm $ exists as real number and the limit $\lim_{E\to +\infty }s_\pm (\lambda ,E) =\ell _{+\infty }^\pm $ exists and belongs to $(0,+\infty ]$ (resp. $[-\infty ,0) $). Moreover, $$ -\infty \leq \ell_{+\infty }^-<\ell_0^-\leq 0\leq \ell_0^+<\ell _{+\infty }^+\leq +\infty . $$ Let us observe that, for any fixed $p>1$ and $\lambda >-m_\pm $, the function $(E,s) \mapsto G_\pm (E,s) $ is continuous on $[0,+\infty ) ^2$ (resp. $[0,+\infty ) \times (-\infty ,0]$) and the function $E\longmapsto s_\pm (\lambda ,E) $ is continuous on $(0,+\infty ) $ and satisfies (\ref{Eq.2}). So, by passing to the limit in (\ref {Eq.2}) as $E$ tends to $0^+$ one obtains $$ 0=\lim_{E\to 0^+}G_\pm (E,s_\pm (\lambda ,E) ) =G_\pm (0,\ell_0^+) \,. $$ Hence, $\ell_0^\pm $ is a zero, belonging to $[0,+\infty ) $ (resp. $(-\infty ,0]$), to the equation in $s$: $$ G_\pm (0,s) =0\mbox{.} $$ By solving this equation one gets: $\ell_0^\pm =0$. Assume that $\ell_{+\infty }^\pm $ is finite, then by passing to the limit in (\ref{Eq.2}) as $E$ tends to $+\infty $ one gets, $$ +\infty =p'(F(\ell_{+\infty }^\pm ) +\lambda \Phi_q(\ell_{+\infty }^\pm ) ) <+\infty , $$ which is impossible. So, $\ell_{+\infty }^\pm =\pm \infty $. \paragraph{Proof of (iii).} Dividing equation (\ref{Eq.2}) by $| s_{\pm }(\lambda ,E) | ^q$ one gets, $$ \frac{E^p}{| s_\pm (\lambda ,E) | ^q}=p^{\prime }(\frac{F(s_\pm (\lambda ,E) ) }{| s_\pm (\lambda ,E) | ^q}+\frac \lambda q) , $$ and by passing to the limit as $E$ tends to $0^+$, (using l'Hospital's rule), $$ \lim_{E\to 0^+}\frac{E^p}{| s_\pm (\lambda ,E) | ^q}=\frac{p'}q(\lim_{E\to 0^+}\frac{f(s_\pm (\lambda ,E) ) }{\varphi_q( s_\pm (\lambda ,E) ) }+\lambda ) =\frac{p'}q(a_0+\lambda ) . $$ The second limit is obtained by the same way. \paragraph{Proof of (iv).} Dividing equation (\ref{Eq.2}) by $| s_{\pm }(\lambda ,E) | ^p$ one gets, \begin{eqnarray*} \lim\limits_{E\to 0^+}\frac{E^p}{| s_\pm (\lambda ,E) | ^p} & = & \lim\limits_{E\to 0^+}p^{\prime }(\frac{F(s_\pm (\lambda ,E) ) }{| s_\pm (\lambda ,E) | ^p}+\frac \lambda q| s_{\pm }(\lambda ,E) | ^{q-p}) \\ & = & \lim\limits_{E\to 0^+}p'| s_\pm ( \lambda ,E) | ^{q-p}(\frac{F(s_\pm ( \lambda ,E) ) }{| s_\pm (\lambda ,E) | ^q}+\frac \lambda q) \\ & = & \frac{p'}q(a_0+\lambda ) \lim\limits_{E\to 0^+}| s_\pm (\lambda ,E) | ^{q-p}. \end{eqnarray*} Therefore, the first limit follows. The second one is obtained by the same way. \paragraph{Proof of (v).} Using l'Hospital's rule one gets, for any $t>0$, \begin{eqnarray*} \lim \limits_{E\to 0^+}\frac{F(s_\pm (\lambda ,E) t) }{E^p} & = & \lim \limits_{E\to 0^+} \frac{t\frac{ds_\pm }{dE}(\lambda ,E) f(s_\pm ( \lambda ,E) t) }{pE^{p-1}} \\ & = & \lim \limits_{E\to 0^+} \frac{t(p-1) E^{p-1}f(s_\pm (\lambda ,E) t) }{(f( s_\pm (\lambda ,E) ) +\lambda \varphi_q(s_{\pm }(\lambda ,E) ) ) pE^{p-1}} \\ & = & \frac t{p'}\lim \limits_{E\to 0^+} \frac{\frac{f(s_\pm (\lambda ,E) t) }{\varphi_q( s_\pm (\lambda ,E) t) }}{(\frac{f(s_{\pm }(\lambda ,E) ) }{\varphi_q(s_\pm ( \lambda ,E) ) }\frac 1{\varphi_q(t) }+\frac \lambda {\varphi_q(t) }) } \\ & = & \frac{t^q}{p'}(\frac{a_0}{a_0+\lambda }) . \end{eqnarray*} The second limit may be computed by the same way, which completes the proof of Lemma \ref{lemma1}. \hfill$\diamondsuit$\smallskip Now, for any $p>1$, $\lambda >-m_\pm $ and $E>0$ we compute $X_\pm (\lambda ,E) $ as defined in Section \ref {sec3}. In fact, for all $E>0, $$$ X_{+}(\lambda ,E) =(0,s_{+}(\lambda ,E) ) ,X_{-}(\lambda ,E) =(s_{-}(\lambda ,E) ,0) , $$ where $s_\pm (\lambda ,E) $ is defined in Lemma \ref{lemma1}. Then $$ r_\pm (\lambda ,E) =s_\pm (\lambda ,E) \mbox{ \ for all }E>0. $$ Hence, for any $p>1,\lambda >-m_\pm $, $0<| s_\pm (\lambda ,E) | <+\infty$ if and only if $E>0$. And for all $E>0$, $$ \pm (f(r_\pm (\lambda ,E) ) +\lambda \varphi_q(r_\pm (\lambda ,E) ) ) = \pm \varphi_q(r_\pm (\lambda ,E) ) ( \frac{f((r_\pm (\lambda ,E) ) ) }{\varphi_q(r_\pm (\lambda ,E) ) }+\lambda ) >0. $$ So, $D_\pm (\lambda ) =(0,+\infty )$ for all $\lambda >-m_\pm $ and $$ D(\lambda ) =D_{+}(\lambda ) \cap D_{-}( \lambda ) =(0,+\infty ) ,\forall \lambda >-m. $$ Before going further in the investigation, from Lemma \ref{lemma1}, we deduce that for any fixed $p>1$ and $\lambda >-m_\pm $, \begin{equation} \pm \frac{\partial r_\pm }{\partial E}(\lambda ,E) =\frac{\pm (p-1) E^{p-1}}{f(r_\pm (\lambda ,E) ) +\lambda \varphi_q(r_\pm (\lambda ,E) ) }>0,\forall E>0. \label{A} \end{equation} \begin{equation} \lim\limits_{E\to 0^+}r_\pm (\lambda ,E) =0\mbox{\ \ \ \ and\ \ \ \ }\lim\limits_{E\to +\infty }r_\pm (\lambda ,E) =\pm \infty \label{B} \end{equation} \begin{equation} \lim\limits_{E\to 0}\frac{E^p}{| r_\pm (\lambda ,E) | ^q}=\frac{p'}q(a_0+\lambda ) ,\lim\limits_{E\to +\infty }\frac{E^p}{| r_\pm ( \lambda ,E) | ^q}=\frac{p'}q(a_\pm +\lambda ) . \label{C} \end{equation} \begin{equation} \label{D} \lim \limits_{E\to 0}\frac E{| r_\pm (\lambda ,E) | }=\left\{ \begin{array}{lcc} 0 & \mbox{if} & q-p>0 \\ (\frac{a_0+\lambda }{p-1}) ^{\frac 1p} & \mbox{if} & q-p=0 \\ +\infty & \mbox{if} & q-p<0 \end{array} \right. \end{equation} \begin{equation} \label{E} \lim \limits_{E\to +\infty }\frac E{| r_\pm (\lambda ,E) | }=\left\{ \begin{array}{lcc} +\infty & \mbox{if} & q-p>0 \\ (\frac{a_\pm +\lambda }{p-1}) ^{\frac 1p} & \mbox{if} & q-p=0 \\ 0 & \mbox{if} & q-p<0 \end{array} \right. \end{equation} \begin{equation} \label{F} \lim \limits_{E\to 0}\frac{F(r_\pm (\lambda ,E) t) }{E^p}=\frac{t^p}{p'}\frac{a_0}{a_0+\lambda },\forall t>0 \end{equation} \begin{equation} \label{G} \lim \limits_{E\to +\infty }\frac{F(r_\pm (\lambda ,E) t) }{E^p}=\frac{t^p}{p'}\frac{a_\pm }{a_{\pm }+\lambda },\forall t>0. \end{equation} At present we define, for any $p>1,$\ $\lambda >-m_\pm $, and $E>0$, the time map, $$ T_\pm (\lambda ,E) :=\pm \int_0^{r_\pm (\lambda ,E) }\left\{ E^p-p'(F(\xi ) +\lambda \Phi _q(\xi ) ) \right\} ^{-\frac 1p}d\xi ,E>0, $$ and a simple change of variables shows that, \begin{equation} T_\pm (\lambda ,E) =| r_\pm (\lambda,E) | \int_0^1\left\{ E^p-p'(F( r_\pm (\lambda ,E) \xi ) +(\lambda /q) | r_\pm (\lambda ,E) \xi | ^q) \right\}^{-\frac 1p}d\xi , \label{U} \end{equation} which may be written as, \begin{eqnarray}\label{H} T_\pm (\lambda ,E) &=&(| r_\pm (\lambda,E) | /E) \int_0^1\big\{ 1-p'(F(r_\pm (\lambda ,E) \xi ) /E^p \\ &&\hspace{3cm}+(\lambda \xi ^q/q) (| r_\pm (\lambda ,E) | ^q/E^p) ) \big\}^{-1/p}d\xi \,. \nonumber \end{eqnarray} Also, we define, the time maps $$ \displaylines{ T_{2n}^\pm (\lambda ,E) :=n(T_{+}(\lambda ,E) +T_{-}(\lambda ,E) ) ,E>0,\lambda >-m_{2n}^\pm ,n\geq 1\,, \cr T_{2n+1}^\pm (\lambda ,E) :=T_{2n}^\pm (\lambda ,E) +T_\pm (\lambda ,E) ,E>0,\lambda >-m_{2n+1}^\pm ,n\geq 0\,. }$$ To prove Theorems \ref{thm1},\ \ref{thm2}, \ref{thm3}, it suffices to compute the limits of these time maps as $E$ tends to $0^+$ and $+\infty $, and then apply the intermediate value theorem. Recall that we have defined in Proposition \ref{prop1} the functions $h_k^\pm $ and let us now define, $$ g_k(\lambda ) =\frac{\lambda_k^{1/p}}{(a_0+\lambda ) ^{1/p}},\mbox{ for all }\lambda >-a_0,\ k\geq 1\,. $$ \begin{lemma} \label{lemma22}Assume that $p,q>1$, then for all $k\geq 1$, and all $\lambda >-m_k^\pm $, \begin{enumerate} \item $\lim\limits_{E\to 0^+}T_k^\pm (\lambda ,E) =+\infty $ and $\lim\limits_{E\to +\infty }T_k^{\pm }(\lambda ,E) =0$, if $q>p$, \item $\lim\limits_{E\to 0^+}T_k^\pm (\lambda ,E) =0$ and $\lim\limits_{E\to +\infty }T_k^\pm ( \lambda ,E) =+\infty $, if $qh_k^{\pm }(\lambda ) $ for all $\lambda >-a_0$. \item[(ii)] If $a_0>c_k^\pm $, then $g_k(\lambda ) -b_k^\pm $. \item[(iii)] If $a_{-}-a_{-}$. \item[(iv)] If $a_{-}h_k^\pm (\lambda ) \mbox{ for all }\lambda \in (\tilde \lambda_k^\pm ,+\infty ) . \end{array} $$ Moreover, the function $a_0\mapsto \tilde \lambda_k^\pm (a_0) $ is strictly increasing on $(a_{-},a_k^\pm ) $ and $$ \lim_{a_0\to a_{-}}\tilde \lambda_k^\pm (a_0) =-a_{-},\mbox{and}\lim_{a_0\to a_k^\pm }\tilde \lambda_k^\pm (a_0) =+\infty . $$ \end{description} \end{lemma} \paragraph{Proof.} It is immediate to prove Assertions {\bf (i)} and {\bf (ii)} of this lemma since the function $a\mapsto (a+\lambda ) ^{-1/p}$ is strictly decreasing. Let us prove assertions {\bf (iii) }and {\bf (iv)}. We are concerned by the case $k=2n+1$ for some $n\geq 1 $ and the superscript is $+:(k=2n+1,+) $. All the remaining cases may be handled similarly. Let $y_{2n+1}^+$ be the function defined on $(-a_{-},+\infty ) $ by $$ y_{2n+1}^+(\lambda ) =(n+1) (a_{+}+\lambda ) ^{-1/p}+n(a_{-}+\lambda ) ^{-1/p}-(2n+1) (a_0+\lambda ) ^{-1/p}. $$ One has, $y_{2n+1}^+(\lambda ) <0$ if and only if $$ ((\frac{n+1}{2n+1}) (a_{+}+\lambda ) ^{-1/p}+(\frac n{2n+1}) (a_{-}+\lambda ) ^{-1/p}) ^p<(a_0+\lambda ) ^{-1} $$ and $(y_{2n+1}^+) '(\lambda ) <0$ if and only if $$ ((\frac{n+1}{2n+1}) (a_{+}+\lambda ) ^{-1-1/p}+(\frac n{2n+1}) (a_{-}+\lambda ) ^{-1-1/p}) ^{p/(p+1) }> (a_0+\lambda ) ^{-1}. $$ Since the function $t\mapsto t^p$ (resp. $t\mapsto -t^{p/(p+1) }$) is convex on $(-a_{-},+\infty ) $, \begin{eqnarray*} \lefteqn{ ((\frac{n+1}{2n+1}) (a_{+}+\lambda ) ^{-1/p}+(\frac n{2n+1}) (a_{-}+\lambda )^{-1/p}) ^p } \\ &&< (\frac{n+1}{2n+1}) (a_{+}+\lambda ) ^{-1}+(\frac n{2n+1}) (a_{-}+\lambda ) ^{-1} \end{eqnarray*} (resp. \begin{eqnarray*} \lefteqn{ ((\frac{n+1}{2n+1}) (a_{+}+\lambda ) ^{-1-1/p}+(\frac n{2n+1}) (a_{-}+\lambda ) ^{-1-1/p}) ^{p/(p+1)} }\\ && >(\frac{n+1}{2n+1}) (a_{+}+\lambda ) ^{-1}+(\frac n{2n+1}) (a_{-}+\lambda ) ^{-1}\mbox{)}. \end{eqnarray*} So, if we define on $(-a_{-},+\infty ) $ the function $x_{2n+1}^+$ by $$ x_{2n+1}^+(\lambda ) =(\frac{n+1}{2n+1}) ( a_{+}+\lambda ) ^{-1}+(\frac n{2n+1}) (a_{-}+\lambda ) ^{-1}-(a_0+\lambda ) ^{-1}, $$ it follows that for all $\lambda >-a_{-}, $$$ x_{2n+1}^+(\lambda ) \leq 0\Longrightarrow y_{2n+1}^+( \lambda ) <0\mbox{\ \ \ and\ \ \ }x_{2n+1}^+(\lambda ) \geq 0\Longrightarrow (y_{2n+1}^+) '(\lambda ) <0. $$ Some simple computations show that for all $\lambda >-a_{-}$ and $\kappa =+,-$, one has, $$ \kappa x_{2n+1}^+(\lambda ) >0\Longleftrightarrow \kappa (\lambda (a_0-a_{2n+1}^+) -( a_{-}a_{+}-a_0a_{2n+1}^-) ) >0. $$ Also, for all $\lambda \in {\mathbb R}$ and $\kappa =+,-$, one has in the case where $a_0>a_{2n+1}^+, $$$ \kappa (\lambda (a_0-a_{2n+1}^+) -( a_{-}a_{+}-a_0a_{2n+1}^-) ) >0\Longleftrightarrow \kappa (\lambda -\Lambda_{2n+1}^+) >0, $$ and in the case where $a_00\Longleftrightarrow \kappa (\lambda -\Lambda_{2n+1}^+) <0. $$ On the other hand one has, $$ a_0-a_{-}\mbox{\ \ \ and\ \ \ }a_0>a_{2n+1}^+\Longrightarrow \Lambda_{2n+1}^+<-a_{-}. $$ Hence, an easy compilation of the above assertions shows that if $a_0-a_{-}) &\Longrightarrow& x_{2n+1}^+(\lambda ) \leq 0 \\ -a_{-}<\lambda <\Lambda_{2n+1}^+&\Longrightarrow& x_{2n+1}^+( \lambda ) <0 \end{eqnarray*} and if $a_0>a_{2n+1}^+$, then $\lambda >-a_{-}$ implies $x_{2n+1}^+(\lambda ) >0$. It remains to study the particular case $a_0=a_{2n+1}^+$. For $\lambda >-a_{-}$, define $$ \psi_\lambda (t) =(\lambda +t) ^{-1},\mbox{ for all }t>-\lambda . $$ Let us observe that $x_{2n+1}^+(\lambda ) >0$ if and only if $$ \psi_\lambda (a_0) <(\frac n{2n+1}) \psi_\lambda (a_{-}) +(\frac{n+1}{2n+1}) \psi_\lambda ( a_{+}) . $$ But since $\psi_\lambda $ is strictly convex, one has \begin{eqnarray*} \psi_\lambda (a_0) =\psi_\lambda (a_{2n+1}^+) &=&\psi_\lambda ((\frac n{2n+1}) a_{-}+(\frac{n+1}{2n+1}) a_{+}) \\ &<& (\frac n{2n+1}) \psi_\lambda (a_{-}) +( \frac{n+1}{2n+1}) \psi_\lambda (a_{+}) , \end{eqnarray*} that is, if $a_0=a_{2n+1}^+$, then $$ \lambda >-a_{-}\Longrightarrow x_{2n+1}^+(\lambda ) >0. $$ Thus, in the case where $a_0\geq a_{2n+1}^+$, $y_{2n+1}^+$ is strictly decreasing on $(-a_{-},+\infty ) $ and by $\lim_{\lambda \to +\infty }y_{2n+1}^+(\lambda ) =0$ it follows that $y_{2n+1}^+$ is strictly positive on $(-a_{-},+\infty ) $. Thus, Assertion {\bf (iii)} is proved. In the case where $a_00$, and $$ \frac{\partial y_{2n+1}^+}{\partial a_0}(a_0,\tilde \lambda _{2n+1}^+(a_0) ) =\frac{2n+1}p(a_0+\tilde \lambda _{2n+1}^+(a_0) ) ^{-1-1/p}>0. $$ So, the function $a_0\mapsto \tilde \lambda_{2n+1}^+(a_0) $ is strictly increasing on $(a_{-},a_{2n+1}^+) $. On the other hand, one has $$ -a_{-}<\tilde \lambda_{2n+1}^+(a_0) <\Lambda _{2n+1}^+(a_0) ;\forall a_0\in (a_{-},a_{2n+1}^+) , $$ and $\lim\limits_{a_0\mapsto a_{-}}\Lambda _{2n+1}^+(a_0) =-a_{-}$, which is easy to check. Thus, $\lim\limits_{a_0\to a_{-}}\tilde \lambda_{2n+1}^+( a_0) =-a_{-}$. Assume that the function $a_0\mapsto \tilde \lambda _{2n+1}^+(a_0) $ is bounded as $a_0$ tends to $a_{2n+1}^+$. Denote by $\lambda_{*}$ its limit, that is, $\lim\limits_{a_0\to a_{2n+1}^+}\tilde \lambda_{2n+1}^+(a_0) =\lambda_{*}$. Since $$ y_{2n+1}^+(a_0,\tilde \lambda_{2n+1}^+(a_0) ) =0,\forall a_0\in (a_{-},a_{2n+1}^+) , $$ it follows that $y_{2n+1}^+(a_{2n+1}^+,\lambda_{*}) =0$, that is, \begin{eqnarray*} \lefteqn{ ((\frac{n+1}{2n+1}) (a_{+}+\lambda_{*}) +(\frac n{2n+1}) (a_{-}+\lambda_{*}) ) ^{-1/p} } \\ &=& (\frac{n+1}{2n+1}) (a_{+}+\lambda_{*}) ^{-1/p}+(\frac n{2n+1}) (a_{-}+\lambda_{*}) ^{-1/p}\,. \end{eqnarray*} By the strict convexity of the function $t\to t^{-1/p}$ on $(0,+\infty ) $, it follows that $a_{+}+\lambda _{*}=a_{-}+\lambda_{*}$, that is $a_{-}=a_{+}$, which contradicts the hypothesis $a_{-}-m_k^\pm$ $$ \lim_{E\to 0^+}T_k^\pm (\lambda ,E) =\frac 12g_k(\lambda ) \quad\mbox{and}\quad \lim_{E\to +\infty }T_k^\pm (\lambda ,E) =\frac 12h_k^\pm ( \lambda ) . $$ On the other hand, for all $\lambda >-a_0:g_k(\lambda ) >h_k^\pm (\lambda ) $. So, since the function $\lambda \mapsto g_k(\lambda ) $ (resp. $\lambda \mapsto h_k^\pm (\lambda ) $) is strictly decreasing on $(-a_0,+\infty ) $ (resp. on $(-b_k^\pm ,+\infty ) $), and $$ g_k(\lambda_k-a_0) =1\mbox{ (resp. }h_k^\pm ( \lambda_k^\pm ) =1\mbox{)}\,, $$ then $$ g_k(\lambda ) >1\mbox{ \ \ if and only if \ \ }-a_0<\lambda <\lambda_k-a_0 $$ (resp. $$ h_k^\pm (\lambda ) <1\mbox{\ \ \ if and only if\ \ \ }\lambda _k^\pm <\lambda \mbox{).} $$ Hence, one has, $$ g_k(\lambda ) >1>h_k^\pm (\lambda ) \mbox{\ \ \ if and only if \ \ }\max (-a_0,\lambda_k^\pm ) <\lambda <\lambda_k-a_0. $$ Then the equation in $E>0:T_k^\pm (\lambda ,E) =1/2$ admits at least a solution for all $\lambda $ satisfying $$ \lambda >-m_k^\pm \quad\mbox{and}\quad \max (-a_0,\lambda _k^\pm ) <\lambda <\lambda_k-a_0, $$ that is, for all $\lambda $ satisfying $$ \max (-m_k^\pm ,\lambda_1^\pm ) <\lambda <\lambda_k-a_0, $$ since $\max (-a_0,-m_k^\pm ) =-m_k^\pm $. If $a_0>c_k^\pm $ then for all $k\geq 1$ and $\lambda >-m_k^\pm $, $$ \lim_{E\to 0^+}T_k^\pm (\lambda ,E) =\frac 12g_k(\lambda ) \quad\mbox{and}\quad \lim_{E\to +\infty }T_k^\pm (\lambda ,E) =\frac 12h_k^\pm ( \lambda ) . $$ On the other hand, for all $\lambda \in (-b_k^\pm ,+\infty )$, $g_k(\lambda ) 1\mbox{\ \ \ if and only if\ \ \ }-b_k^\pm <\lambda <\lambda_k^\pm \mbox{).} $$ Hence, one has, $$ g_k(\lambda ) <10:T_k^\pm (\lambda ,E) =1/2$ admits at least a solution for all $\lambda $ satisfying $$ \lambda >-m_k^\pm \quad\mbox{and}\quad \max (-b_k^{\pm },\lambda_k-a_0) <\lambda <\lambda_k^\pm , $$ that is, for all $\lambda $ satisfying $$ \max (-m_k^\pm ,\lambda_k-a_0) <\lambda <\lambda_k^\pm $$ since $\max (-b_k^\pm ,-m_k^\pm ) =-m_k^\pm $. The proof of Theorem \ref{thm1} is complete. \hfill$\diamondsuit$\smallskip \paragraph{Proof of Theorem \ref{thm2}.} If $a_{-}-m_\pm $, $$ \lim_{E\to 0^+}T_1^\pm (\lambda ,E) =\frac 12g_1(\lambda ) \quad\mbox{and}\quad \lim_{E\to +\infty }T_1^\pm (\lambda ,E) =\frac 12h_1^\pm ( \lambda ) . $$ On the other hand, one has, $$ \mbox{for all }\lambda >-a_0,g_1(\lambda ) >h_1^+( \lambda ) \mbox{\ \ \ \ (resp. for all }\lambda >-a_{-},g_1( \lambda ) -m_k^\pm =-m$, $$ \lim_{E\to 0^+}T_k^\pm (\lambda ,E) =\frac 12g_k(\lambda ) \quad\mbox{and}\quad \lim_{E\to +\infty }T_k^\pm (\lambda ,E) =\frac 12h_k^\pm ( \lambda ) . $$ On the other hand, one has for all $\lambda >-a_{-},g_k(\lambda ) 0:T_k^{\pm }(\lambda ,E) =1/2$ admits at least a solution for all $\lambda $ satisfying $$ \max (-m,\lambda_k-a_0) <\lambda <\lambda_k^\pm . $$ If $a_{-}-m$, $$ \lim_{E\to 0^+}T_k^\pm (\lambda ,E) =\frac 12g_k(\lambda ) \quad\mbox{and}\quad \lim_{E\to +\infty }T_k^\pm (\lambda ,E) =\frac 12h_k^\pm ( \lambda ) . $$ Since the function $a_0\mapsto \tilde \lambda_k^\pm (a_0) \in (-a_{-},\Lambda_k^\pm ) $ is strictly increasing on the interval $(a_{-},a_k^\pm ) $ then the function $a_0\mapsto h_k^\pm (\tilde \lambda_k^\pm (a_0) ) $ is strictly decreasing on $(a_{-},a_k^\pm ) $. Also, one has, $$ \lim_{a_0\to a_{-}}h_k^\pm (\tilde \lambda_k^\pm ( a_0) ) =\lim_{x\to -a_{-}}h_k^\pm (x) =+\infty , $$ and $$ \lim_{a_0\to a_k^\pm }h_k^\pm (\tilde \lambda_k^{\pm }(a_0) ) =\lim_{x\to +\infty }h_k^\pm ( x) =0\,. $$ Then there exists a unique $\tilde a_k^\pm \in (a_{-},a_k^{\pm }) $ such that for all $a_0\in (a_{-},a_k^\pm ) $ one gets, \begin{eqnarray*} h_k^\pm (\tilde \lambda_k^\pm (a_0) ) >1 &\Longleftrightarrow& a_{-}\tilde \lambda_k^\pm (a_0) $, $$ h_k^\pm (\lambda ) <10:T_k^\pm (\lambda ,E) =1/2$ admits at least a solution for all $\lambda $ satisfying $$ \lambda >-m\quad\mbox{and}\quad \lambda_k^\pm <\lambda <\lambda_k-a_0\,, $$ that is, for all $\lambda $ satisfying, $\max (-m,\lambda_k^\pm ) <\lambda <\lambda_k-a_0$. Case where $\tilde a_k^\pm \tilde \lambda_k^\pm (a_0)$, $$ h_k^\pm (\lambda ) 0:T_k^\pm (\lambda ,E) =1/2$ admits at least a solution for all $\lambda $ satisfying $$ \lambda >-m\mbox{\ \ \ \ and\ \ \ }\max (-a_{-},\lambda _k-a_0) <\lambda <\lambda_k^\pm , $$ that is, for all $\lambda $ satisfying $\max (-m,\lambda _k-a_0) <\lambda <\lambda_k^\pm $, this is so because $\max ( -m,-a_{-}) =-m$. The proof of Theorem \ref{thm2} is complete. \hfill$\diamondsuit$ \paragraph{Proof of Theorem \ref{thm3}.} If $q-p>0$ (resp. $q-p<0$), by Lemma \ref{lemma22}, for all $\lambda >-m_k^\pm ,k\geq 1, $$$ \lim_{E\to 0^+}T_k^\pm (\lambda ,E) =+\infty \mbox{ (resp. }=0\mbox{) and }\lim_{E\to 0^+}T_k^\pm (\lambda ,E) =0\mbox{ (resp. }=+\infty \mbox{).} $$ So, the intermediate value theorem implies that the equation in $E>0:T_k^\pm (\lambda ,E) =1/2$ admits at least a solution for all $\lambda >-m_k^\pm $. Theorem \ref{thm3} is proved. \hfill$\diamondsuit$\smallskip \paragraph{Remark 1.} Some easy computations show that $$ \frac{\partial T_{+}}{\partial E}(\lambda ,E) =\frac{(p') ^{-\frac1p}}p (\frac{\partial r_{+}}{\partial E}(\lambda ,E) ) \int_0^1\frac{(H(\lambda ,r_{+}(\lambda ,E) ) -H(\lambda ,r_{+}(\lambda ,E) \xi ) ) }{(F(\lambda ,r_{+}(\lambda ,E) ) -F(\lambda ,r_{+}(\lambda ,E) \xi ) ) ^{1+\frac 1p}}d\xi , $$ where $H(\lambda ,x) :=p\tilde F(\lambda ,x) -x\tilde f(\lambda ,x) ,\tilde F(\lambda ,x) =\lambda \Phi_q(x) +F(x) $ and $\tilde f( \lambda ,x) =\lambda \varphi_q(x) +f(x) $. So, if $q=p$, one gets, $$ \frac d{dx}(\frac{f(x) }{\varphi_p(x) }) =\frac{-1}{x\varphi_p(x) }\frac{\partial H}{\partial x}(\lambda ,x) ,x>0. $$ Hence, in the particular case where the function $x\mapsto f(x) /\varphi_p(x) ,(q=p) $ is strictly decreasing on $(0,+\infty ) $ the function $x\mapsto H(\lambda ,x) $ is strictly increasing on $(0,+\infty ) $ and then the time map $E\mapsto T_{+}(\lambda ,E) $ is strictly increasing on $(0,+\infty ) $. So, uniqueness (if existence) of the solution of the equation in $E>0:T_{+}(\lambda ,E) =1/2$ is guaranteed. The exact number of positive solution(s) in $A_1^+$ should be obtained with this additional condition. The same remark works for the result of Guedda and Veron \cite{GueddaVeron}. That is, their Theorem 2.1 holds without there condition (2.7). \paragraph{Remark 2.} If $f$ is an odd function, $a_{-}=a_{+}$. The statements of Theorems \ref{thm1}, \ref{thm2}, and \ref{thm3}\ may be simplified in this particular case. \paragraph{Remark 3.} Several corollaries may be deduced from Theorems \ref{thm1},\ \ref {thm2},\ and \ref{thm3} and the above remarks. In fact, one may draw some bifurcation diagrams and compute a lower bound on the number of solutions of problem (\ref{p1}) in some cases and the exact number of positive solutions in others. This would require more space and patience, and is left to the diligent, patient reader. However, a qualitative feature of the variations of the bifurcation branches as $a_0$ varies is known. In fact, if $a_{-}\tilde a_k^\pm $. The case where $a_0=\tilde a_k^\pm $ remains an open question. If $a_{-}>a_{+}$, one may study the asymmetric case, $a_{-}>a_0>a_{+}$, as in Theorem \ref{thm2}. \paragraph{Remark 4.} One may observe that a common feature in Theorems \ref{thm1},\ \ref {thm2},\ and \ref{thm3} is that the parameter $\lambda $ is taken, in particular, such that the function $x\mapsto \lambda +f(x) /\varphi_q(x) $ is strictly positive on $(-\infty ,0) $ and/or $(0,+\infty ) $. Some cases where this function changes sign once are studied by Guedda and Veron \cite{GueddaVeron} and by Boucherif, Bouguima and Derhab \cite{BBD}, both for the particular case where $f$ is odd. So, it is reasonable to ask the question of what happens if $f$ is not necessarily odd. \section{Appendix: Historical overview on time maps} At the beginning of time maps' history, the authors used them with the one dimensional Laplacian operator, that is, with the one dimensional p-Laplacian and $p=2$. From the 1960's we can mention Opial \cite{Opial1}, \cite {Opial2}, Urabe \cite{Urabe1}, \cite{Urabe2}, \cite{Urabe3}, \cite{Urabe4}, \cite{Urabe5}, Pimbley \cite{Pimbley1}, \cite{Pimbley2}, and Gavalas \cite {Gavalas}. In the early 1970's, Laetsch \cite{Laetsch} used time maps to study positive solutions to a class of boundary-value problems with Dirichlet boundary data. Since then many authors have referred to his work. We also want to mention Chafee and Infante \cite{ChafeeInfante}, and Chafee \cite{Chafee}. Brown and Budin \cite{BrownBudin1}, \cite{BrownBudin2} used the time map approach to study positive solutions to some boundary-value problems. Independently and about the same time, De-Mottoni and Tesei \cite{DeMottoniTesei1} studied positive solutions of some other class of boundary-value problems by means of the same method. In the early 1980's Smoller and Wasserman \cite{SmollerWasserman} introduced a technique that, in some circumstances, can be used to prove uniqueness of the critical point of time maps. Their technique has been used subsequently by many authors; see for instance, Ammar Khodja \cite{AmmarKhodja}, Ramaswamy \cite{Ramaswamy}, S. H. Wang and Kazarinoff, \cite{SHWang1}, \cite{SHWang2}, S. H. Wang and F. P. Lee \cite{SHWang3}, S. H. Wang \cite{SHWang4}, \cite{SHWang5}, \cite{SHWang6}, and recently by Addou and Benmeza\"\i\ \cite{add3}. The study of sign-changing solutions by means of time maps was initiated by De-Mottoni and Tesei \cite{DeMottoniTesei2}, and independently, some years after, by Shivaji \cite{Shivaji2}. During the last two decades, time maps have been used in many publications. Besides the above mentioned papers, we want to add the following ones: Addou and Ammar Khodja \cite{Addou1}, Anuradha, Shivaji and Zhu \cite {Anuradha1}, \cite{Anuradha15}, Anuradha and Shivaji \cite{Anuradha2}, Anuradha, C. Brown and Shivaji \cite{Anuradha3}, Brown, Ibrahim and Shivaji \cite{Brown}, Brunovsky and Chow \cite{Brunovsky}, Castro and Shivaji \cite {Castro}, \cite{Castro2}, Ding and Zanolin \cite{Ding1}, \cite{Ding2}, Fernandes \cite{Fernandes}, Fonda and Zanolin \cite{Fonda}, Fonda, Gossez and Zanolin \cite{Fonda2}, Schaaf \cite{Schaaf}, Shivaji \cite{Shivaji1},\ \cite{Shivaji2},\ \cite{Shivaji3}, Smoller, Tromba and Wasserman \cite {Smoller}, Smoller and Wasserman \cite{SmollerWasserman1}, \cite {SmollerWasserman2}, \cite{SmollerWasserman3}. Notice that this list is in alphabetical order, and is not complete by any means. The differential operator in the equations studied in these papers is the $p$-Laplacian with $p=2$. For more general differential operators, see references listed in Section 3 of this paper. \paragraph{Acknowledgment.} I want to thank Professor Julio G. Dix for his help in improving the presentation of an early version of this article. \begin{thebibliography}{99} \bibitem{Addou1} {\sc Addou, I., and F. Ammar Khodja}, {\it On the number of solutions of a nonlinear boundary value problem,} C.R.A.S., Paris, {\bf t. 321} S\'erie I, (1995), pp. 409-412. \bibitem{add1} {\sc Addou, I., S. M. Bouguima, M. Derhab, and Y. S. Raffed,}{\it \ On the number of solutions of a quasilinear elliptic class of B.V.P. with jumping nonlinearities, }Dynamic Syst. Appl. {\bf 7} (1998), pp. 575-599. \bibitem{add2} {\sc Addou, I., and A. Benmeza\"\i ,}{\it \ Boundary value problems for the one dimensional p-Laplacian with even superlinearity, }Electron. J. Diff. Eqns. {\bf 1999} (1999), No. 09, pp. 1-29. \bibitem{add3} {\sc Addou, I., and A. Benmeza\"\i ,}{\it \ Exact number of positive solutions for a class of quasilinear boundary value problems,} Dynamic Syst. Appl. {\bf 8} (1999), pp. 147-180. \bibitem{az1} {\sc Amann, H., and E. Zehnder,}{\it \ Nontrivial solutions for a class of non resonance problems and applications to nonlinear differential equations, }Ann. Scuola Norm. Sup. Pisa {\bf 7} (1980), pp. 539-603. \bibitem{AmmarKhodja} {\sc Ammar Khodja, F.}{\it , }Th\`ese 3\`e cycle, Univ. Pierre et Marie Curie, Paris VI, (1983). \bibitem{Anuradha1} {\sc Anuradha, V., R. Shivaji, and J. Zhu}, {\it On S-Shaped bifurcation curves for multiparameter positone problems,} Appl. Math. Comput. {\bf 65} (1994), pp. 171-182. \bibitem{Anuradha15} {\sc Anuradha, V., R. Shivaji, and J. Zhu},{\it \ Positive solutions for classes of positone problems when a parameter is negative,} Dynamic Syst. Appl. {\bf 3} (1994), pp. 235-250. \bibitem{Anuradha2} {\sc Anuradha, V., R. Shivaji}, {\it Sign changing solutions for a class of superlinear multi-parameter semi-positone problems}, Nonlinear Anal. T.M.A. {\bf 24} (1995), pp. 1581-1596. \bibitem{Anuradha3} {\sc Anuradha, V., C. Brown, and R. Shivaji}, {\it Explosive nonnegative solutions to two point boundary value problems}, Nonlinear Anal. T.M.A. {\bf 26} (1996), pp. 613-630. \bibitem{Berestycki} {\sc Berestycki, H.},{\it \ On some nonlinear Sturm-Liouville problems, }J. Diff. Eqns. {\bf 26} (1977), pp. 375-390. \bibitem{Boccardo} {\sc Boccardo, L. P. Dr\'abek, D. Giachetti, and M. Ku\v cera,}{\it \ Generalization of Fredholm alternative for nonlinear differential operators,} Nonlinear Anal. T.M.A. {\bf 10} (1986), pp. 1083-1103. \bibitem{BBD} {\sc Boucherif, A., S.M. Bouguima, and M. Derhab,}{\it \ On the zeros of solutions of a quasilinear elliptic boundary value problems.} Preprint. \bibitem{BrownBudin1} {\sc Brown, K.J., and H. Budin}, {\it Multiple positive solutions for a class of nonlinear boundary value problems}, J. Math. Anal. Appl. {\bf 60} (1977), pp. 329-338. \bibitem{BrownBudin2} {\sc Brown, K.J., and H. Budin}, {\it On the existence of positive solutions for a class of semilinear elliptic boundary value problems}, SIAM J. Math. Anal. {\bf 10} (1979), pp. 875-883. \bibitem{Brown} {\sc Brown, K.J., M.M.A., Ibrahim, and R. Shivaji,} {\it S-shaped bifurcation curves}, Nonlinear Anal. T.M.A. {\bf 5} (1981), pp. 475-486. \bibitem{Brunovsky} {\sc Brunovsky, P., and S.N. Chow}, {\it Generic properties of stationary state solutions of reaction-diffusion equations,} J. Diff. Eqns. {\bf 53} (1984), pp. 1-23. \bibitem{cl} {\sc Castro, A. and A. C. Lazer,}{\it \ Critical point theory and the number of solutions of a nonlinear Dirichlet problem, }Ann. Mat. Pura Appl. {\bf 120} (1979), pp. 113-137. \bibitem{Castro} {\sc Castro, A., and R. Shivaji,} Multiple solutions for a Dirichlet problem with jumping nonlinearities, in Trends in the Theory and Practice of Non-Linear Analysis (V. Lakshmikantham , Ed.) Elsevier Science Publishers B. V., North-Holland, (1985), pp. 97-101. \bibitem{Castro2} {\sc Castro, A., and R. Shivaji,} {\it Non-negative solutions for a class of non-positone problems,} Proc. Roy. Soc. Edin. {\bf 108A} (1988), pp. 291-302. \bibitem{ChafeeInfante} {\sc Chafee, N., and E.F. Infante}, {\it A bifurcation problem for a nonlinear partial differential equation of parabolic type}, Applicable Anal. {\bf 4} (1974), pp. 17-37. \bibitem{Chafee} {\sc Chafee, N.}, {\it Asymptotic behavior for solutions of a one-dimensional parabolic equation with homogenous Neumann boundary conditions}, J. Diff. Eqns. {\bf 18} (1975), pp. 111-134. \bibitem{DelpinoMana1} {\sc Del Pino, M., and R. Man\'asevich,}{\it \ Multiple solutions for the p-Laplacian under global nonresonance}, Proc. Amer. Math. Soc. {\bf 112} (1991), pp. 131-138. \bibitem{DeMottoniTesei1} {\sc De Mottoni, P., and A. Tesei,} {\it On a class of non linear eigenvalue problems}, Boll. U.M.I. {\bf 14-B} (1977), pp. 172-189. \bibitem{DeMottoniTesei2} {\sc De Mottoni, P., and A. Tesei}, {\it On the solutions of a class of nonlinear Sturm-Liouville problems}, SIAM. J. Math. Anal. {\bf 9} (1979), pp. 1020-1029. \bibitem{Ding1} {\sc Ding, T., and F. Zanolin}, {\it Time-maps for the solvability of periodically perturbed nonlinear duffing equations,} Nonlinear Anal. T.M.A. {\bf 17} (1991), pp. 635-653. \bibitem{Ding2} {\sc Ding, T., and F. Zanolin}, {\it Subharmonic solutions of second order nonlinear equations: a time-map approach}, Nonlinear Analysis T. M. A. {\bf 20} (1993), pp. 509-532. \bibitem{Drabek1} {\sc Dr\'abek, P.,}{\it \ Ranges of a-homogeneous operators and their perturbations, \v Cas. p\v est. mat. }{\bf 105} (1980), pp. 167-183. \bibitem{E} {\sc Esteban, M. J.,}{\it \ Multiple solutions of semilinear elliptic problems in a ball, }J. Diff. Eqns. {\bf 57} (1985), pp. 112-137. \bibitem{Fernandes} {\sc Fernandes, M.L.C.}, {\it On an elementary phase-plane analysis for solving second order BVPs,} Ricerche di Matematica, {\bf XXXVII} (1988), pp. 189-202. \bibitem{Fonda} {\sc Fonda, A., and F. Zanolin}, {\it On the use of time-maps for the solvability of nonlinear boundary value problems}, Arch. Math. {\bf 59} (1992), pp. 245-259. \bibitem{Fonda2} {\sc Fonda, A., J.-P. Gossez, and F. Zanolin}, {\it On a nonresonance condition for a semilinear elliptic problem}, Diff. Integral Eqns. {\bf 4} (1991), pp. 945-951. \bibitem{GarciaetAl1} {\sc Garc\'\i a-Huidobro, M., R. Man\'asevich, and F. Zanolin}, Strongly nonlinear second-order ODE's with unilateral conditions, Diff. Integral Eqns. {\bf 6} (1993), pp. 1057-1078. \bibitem{GarciaetAl2} {\sc Garc\'\i a-Huidobro, M., R. Man\'asevich, and F. Zanolin}, On a pseudo Fucik spectrum for strongly nonlinear second-order ODEs and an existence result, J. Comput. Applied Math. {\bf 52} (1994), pp. 219-239. \bibitem{GarciaetAl3} {\sc Garc\'\i a-Huidobro, M., R. Man\'asevich, and F. Zanolin}, A Fredholm-like result for strongly nonlinear second order ODE's, J. Diff. Eqns. {\bf 114} (1994), pp. 132-167. \bibitem{GarciaetAl4} {\sc Garc\'\i a-Huidobro, M., R. Man\'asevich, and F. Zanolin}, Strongly nonlinear second-order ODEs with rapidly growing terms, J. Math. Anal. Appl. {\bf 202 }(1996), pp. 1-26. \bibitem{GarciaetAl5} {\sc Garc\'\i a-Huidobro, M., R. Man\'asevich, and F. Zanolin}, {\it Infinitely many solutions for a Dirichlet problem with a nonhomogeneous p-Laplacian-like operator in a ball}, Advances Diff. Eqns. {\bf 2} (1997), pp. 203-230. \bibitem{GarciaUbilla} {\sc Garc\'\i a-Huidobro, M., and P. Ubilla}{\it , Multiplicity of solutions for a class of nonlinear second-order equations}, Nonlinear Anal. T.M.A. {\bf 28} (1997), pp. 1509-1520. \bibitem{Gavalas} {\sc Gavalas, G.R}., Nonlinear differential equations of chemically reacting systems, Springer Tracts in Natural Philosophy N. {\bf 17}, Springer, New York, (1968). \bibitem{GueddaVeron} {\sc Guedda, M., and L. Veron,}{\it \ Bifurcation phenomena associated to the p-Laplace operator, }Trans. Amer. Math. Soc. {\bf 310 }(1987), pp. 419-431. \bibitem{h} {\sc Hess, P.,}{\it \ On nontrivial solutions of a nonlinear elliptic boundary value problem,} Conf. Sem. Mat. Univ. Bari, {\bf 173} (1980). \bibitem{Huang} {\sc Huang Y. X., and G. Metzen, }{\it The existence of solutions to a class of semilinear differential equations,} Diff. Integral Eqns. {\bf 8} (1995), pp. 429-452. \bibitem{Laetsch} {\sc Laetsch, T.}, {\it The number of solutions of a nonlinear two point boundary value problem}, Indiana Univ. Math. J. {\bf 20} (1970), pp. 1-13. \bibitem{ManaZanolin} {\sc Man\'asevich, R., and F. Zanolin},{\it \ Time-mappings and multiplicity of solutions for the one-dimensional p-Laplacian,} Nonlinear Analysis T. M. A. {\bf 21} (1993), pp. 269-291. \bibitem{ManasevichetAl} {\sc Man\'asevich, R., F. I. Njoku and F. Zanolin}, {\it Positive solutions for the one-dimensional p-Laplacian}, Diff. Integral Eqns. {\bf 8} (1995), pp. 213-222. \bibitem{Opial1} {\sc Opial, Z.}, {\it Sur les solutions p\'eriodiques de l'\'equation diff\'erentielle }$x^{\prime \prime }+g(x) =p( t) $, Bull. Acad. Pol. Sci. S\'erie Sci. Math. Astr. Phys. {\bf 8} (1960), pp. 151-156. \bibitem{Opial2} {\sc Opial, Z.,} {\it Sur les p\'eriodes des solutions de l'\'equation diff\'erentielle }$x^{\prime \prime }+g(x) =0$, Ann. Pol. Math. {\bf 10} (1961), pp. 49-72. \bibitem{Pimbley1} {\sc Pimbley, G.H.}, {\it A sublinear Sturm-Liouville problem}, J. Math. Mech. {\bf 11} (1962), pp. 121-138. \bibitem{Pimbley2} {\sc Pimbley, G.H.}, {\it Eigenfunction branches of nonlinear operators, and their bifurcation, }Lecture Notes Math. {\bf 104} (1969). Springer Verlag. \bibitem{Ramaswamy} {\sc Ramaswamy, M.}{\it , }Th\`ese 3\`e cycle, Univ. Pierre et Marie Curie, Paris VI, (1983). \bibitem{Schaaf} {\sc Schaaf, R.}, Global solution branches of two point boundary value problems, Lecture Notes Math. {\bf 1458} (1990). Springer Verlag. \bibitem{Shivaji1} {\sc Shivaji, R.}, {\it Uniqueness results for a class of positone problems}, Nonlinear Anal. T.M.A. {\bf 7} (1983), pp. 223-230. \bibitem{Shivaji2} {\sc Shivaji, R.}, Perturbed bifurcation theory for a class of autonomous ordinary differential equations, in: Lakshmikantham V. (ed.), Trends in theory and practice of nonlinear differential equations (Marcel Dekker, Inc., New York and Basel, 1984). \bibitem{Shivaji3} {\sc Shivaji, R.}{\it , A comparison of two methods on non-uniqueness for a class of two-point boundary value problems}, Applicable Anal. {\bf 27} (1988), pp. 173-180. \bibitem{SmollerWasserman} {\sc Smoller, J., and A. Wasserman}, {\it Global bifurcation of steady-state solutions}, J. Diff. Eqns. {\bf 39} (1981), pp. 269-290. \bibitem{SmollerWasserman1} {\sc Smoller, J., and A. Wasserman}, {\it % Generic bifurcation of steady state solutions}, J. Diff. Eqns. {\bf 52} (1984), pp. 432-438. \bibitem{SmollerWasserman2} {\sc Smoller, J., and A. Wasserman},{\it \ Existence, uniqueness, and nondegeneracy of positive solutions of semilinear elliptic equations,} Commun. Math. Phys. {\bf 95} (1984), pp. 129-159. \bibitem{SmollerWasserman3} {\sc Smoller, J., and A. Wasserman}, {\it On the monotonicity of the time-map}, J. Diff. Eqns. {\bf 77} (1989), pp. 287-303. \bibitem{Smoller} {\sc Smoller, J., A. Tromba, and A. Wasserman},{\it \ Non-degenerate solutions of boundary-value problems}, Nonlinear Anal. T.M.A. {\bf 4} (1980), pp. 207-215. \bibitem{S} {\sc Struwe, M.,}{\it \ Infinitely many solutions of superlinear boundary value problems with rotational symmetry,} Arch. Math. {\bf 36,} No. 4 (1981), pp. 360-369. \bibitem{Ubilla} {\sc Ubilla, P.}, {\it Multiplicity results for the 1-dimensional generalized p-Laplacian,} J. Math. Anal. Appl. {\bf 190} (1995), pp. 611-623. \bibitem{Urabe1} {\sc Urabe, M.}, {\it Potential forces which yield periodic motions of a fixed period}, J. Math. Mech. {\bf 10} (1961), pp. 569-578. \bibitem{Urabe2} {\sc Urabe, M.}, {\it The potential force yielding a periodic motion whose period is an arbitrary continuous function of the amplitude of the velocity}, Arch. Rational Mech. Anal. {\bf 11} (1962), pp. 27-33. \bibitem{Urabe3} {\sc Urabe, M.}, {\it The potential force yielding a periodic motion whose period is an arbitrary continuously differentiable function of the amplitude}, J. Sci. Hiroshima Univ. Ser. A-{\bf I 26} (1962), pp. 93-109. \bibitem{Urabe4} {\sc Urabe, M.}, {\it The potential force yielding a periodic motion with arbitrary continuous half-periods}, J. Sci. Hiroshima Univ. Ser. A-{\bf I 26} (1962), pp. 111-122. \bibitem{Urabe5} {\sc Urabe, M.}, {\it Relations between periods and amplitudes of periodic solutions of }$x^{\prime \prime }+g(x) =0 $, Funkcialaj Ekvacioj {\bf 6} (1964), pp. 63-88. \bibitem{SHWang1} {\sc Wang, S.H., and N. D. Kazarinoff}{\it , Bifurcation and stability of positive solutions of a two-point boundary value problem,} J. Austral. Math. Soc. Series A {\bf 52} (1992), pp. 334-342. \bibitem{SHWang2} {\sc Wang, S.H., and N. D. Kazarinoff}{\it , Bifurcation of steady-state solutions of a scalar reaction-diffusion equation in one space variable,} J. Austral. Math. Soc. Series A {\bf 52} (1992), pp. 343-355. \bibitem{SHWang3} {\sc Wang, S.H., and F.P. Lee}, {\it Bifurcation of an equation from catalysis theory}, Nonlinear Anal. T.M.A. {\bf 23} (1994), pp. 1167-1187. \bibitem{SHWang4} {\sc Wang, S.H.,}{\it \ On the time map of a nonlinear two point boundary value problem, }Diff. Integral. Eqns{\it . }{\bf 7} (1994), pp. 49-55. \bibitem{SHWang5} {\sc Wang, S.H.}, {\it On S-Shaped bifurcation curves}, Nonlinear Anal. T.M.A. {\bf 22} (1994), pp. 1475-1485. \bibitem{SHWang6} {\sc Wang, S.H.}, {\it Rigorous analysis and estimates of S-shaped bifurcation curves in a combustion problem with general Arrhenius reaction-rate laws}, Proc. Roy. Soc. Lond. {\bf A 453} (1997), pp. 1-18. \bibitem{W} {\sc Wang, J.,}{\it \ The existence of positive solutions for the one dimensional p-Laplacian,} Proc. Amer. Math. Soc. {\bf 125} (1997), pp. 2275-2283. \end{thebibliography} \noindent{\sc Idris Addou}\\ USTHB, Institut de Math\'ematiques \\ El-Alia, B.P. no. 32, Bab-Ezzouar\\ 16111, Alger, Alg\'erie\\ E-mail address: idrisaddou@hotmail.com \end{document}