Congruence modulo p

The name of the pictureThe name of the pictureThe name of the pictureClash Royale CLAN TAG#URR8PPP











up vote
4
down vote

favorite
1












Let $p$ be an odd prime and let $1leq n<p-1.$ Show that $$sum_t=1^pt^n equiv 0 pmodp$$



Remark: It seems one can not apply Fermat's little theorem directly as $n<p-1$










share|cite|improve this question



























    up vote
    4
    down vote

    favorite
    1












    Let $p$ be an odd prime and let $1leq n<p-1.$ Show that $$sum_t=1^pt^n equiv 0 pmodp$$



    Remark: It seems one can not apply Fermat's little theorem directly as $n<p-1$










    share|cite|improve this question

























      up vote
      4
      down vote

      favorite
      1









      up vote
      4
      down vote

      favorite
      1






      1





      Let $p$ be an odd prime and let $1leq n<p-1.$ Show that $$sum_t=1^pt^n equiv 0 pmodp$$



      Remark: It seems one can not apply Fermat's little theorem directly as $n<p-1$










      share|cite|improve this question















      Let $p$ be an odd prime and let $1leq n<p-1.$ Show that $$sum_t=1^pt^n equiv 0 pmodp$$



      Remark: It seems one can not apply Fermat's little theorem directly as $n<p-1$







      elementary-number-theory






      share|cite|improve this question















      share|cite|improve this question













      share|cite|improve this question




      share|cite|improve this question








      edited Oct 31 '12 at 14:15









      Julián Aguirre

      65.3k23894




      65.3k23894










      asked Oct 31 '12 at 12:37









      user31899

      1,5791030




      1,5791030




















          5 Answers
          5






          active

          oldest

          votes

















          up vote
          6
          down vote



          accepted










          We might as well sum to $p-1$. Let $g$ be a primitive root of $p$. Then $1$, $2$, $3$, and so on up to $p-1$ are congruent, in some order, to $g^0$, $g^1$, up to $g^p-2$.
          So modulo $p$ the terms in our sum are congruent to
          $g^0$, $g^n$, $g^2n$, up to $g^(p-2)n$. Sum this geometric progression. We get $dfracg^(p-1)n-1g^n-1$. The numerator is congruent to $0$. There is no problem with the denominator, since $nlt p-1$.



          If one prefers not to use fractions, equivalently note that
          $$left(1+g^n+g^2n+cdots+g^(p-2)nright)(1-g^n)=1-g^(p-1)n.$$
          The right-hand side is congruent to $0$. But $1-g^n$ is not congruent to $0$, and the result follows.






          share|cite|improve this answer





























            up vote
            6
            down vote













            Since the sets of reduced residue classes $1, 2, . . . , (p − 1)$ and $g, 2g, . . . , (p − 1)g$ are the same if $(g,p)=1$ the reason being:
            if $r_1gequiv r_2gpmod p,$ where $1le r_1< r_2le p-1$
            $implies r_1equiv r_2$ which is impossible, so $r_1g≢r_2gpmod p$.



            So, $sum_1le xle p-1 x^n equiv sum_1le yle p-1 (gy)^n pmod p$



            $implies pmid(g^n-1)sum_1le xle p-1 x^n $



            If we take $g$ to be a primitive root, $pmid (g^n-1)iff phi(p)=(p-1)mid n$



            Since $1le n<p-1,p$ can not divide $(g^n-1)implies sum_1le xle p-1 x^nequiv 0pmod p$



            If $(p-1)mid n, sum_1le xle p-1 x^nequiv sum_1le xle p-11pmod p=p-1equiv -1pmod p $






            share|cite|improve this answer





























              up vote
              2
              down vote













              Call
              $$ p_n = sum_k=1^p k^n. $$
              Since, by Fermat's little theorem, $p_nequiv p_n-p+1pmodp$ for all $n > p-1$, it is sufficient to show that
              $$ forall nin[0,p-2],quad p_nequiv 0pmodp, $$
              where the cases $n=0$ and $n=1$ are trivial. Since:
              $$ x(x-1)cdotldotscdot(x-p+1)equiv x^p-xpmodp, $$
              if $e_k(x_1,ldots,x_p)$ is the $k$-th elementary symmetric polynomial, we have:
              $$ forall kin[0,p-2],quad e_k(1,ldots,p)equiv 0pmodp, $$
              so, in virtue of Newton-Girard identities and induction we have:
              $$forall kin[0,p-2],quad p_kequiv 0pmodp, $$
              QED.






              share|cite|improve this answer





























                up vote
                0
                down vote













                EDIT:This is a wrong answer($r^n equiv t^n pmod p not Rightarrow r equiv t pmod p$). See the comments.



                For $n=1$ is correct.



                Now for any $n in 1,2,ldots,p-2$ we have that $1,2,ldots,p-1equiv 1,2^n,ldots,(p-1)^n pmod p$ since $r^n equiv t^n pmod p iff r equiv t pmod p$. Hence $$sum_t=1^pt^n equiv sum_t=1^pt equiv 0 pmodp .$$






                share|cite|improve this answer






















                • I don't thing the $equiv$ is true. For example, if $n=2$, the sequence $1,2^2,3^2,...,(p-1)^2$ only has $fracp-12$ distinct elements
                  – Thomas Andrews
                  Oct 31 '12 at 12:52






                • 1




                  Sorry, but this isn't completely watertight. With $n = 2$ and $p = 5$, you get $1^2, 2^2, 3^2, 4^2 cong 1, 4, 4, 1$
                  – Arthur
                  Oct 31 '12 at 12:53










                • Note that $1^2equiv (p-1)^2pmod p$, but $1notequiv p-1pmod p$
                  – Thomas Andrews
                  Oct 31 '12 at 12:53

















                up vote
                0
                down vote













                WLOG assume $nk = p-1$ (if $n$ had factors coprime to $p-1$ they would just permute the sum, not change it's value, so remove them).



                Put $S = (mathbb Z/pmathbb Z)^times$ and let $S^n$ denote the subgroup $ s in S $. The kernel of the map $S to S^n$ is the set of solutions of $z^n=1$ so $n|S^n| = |S|$. Details about subgroup here



                This tells us that (mod p) the sum is equal to $n$ times the sum of the elements of $S^n$ so we just need to show that $$sum_s in S^ns equiv 0 pmod p$$ but since this is the set of solutions of $z^n-1=0$ by Vieta's formula they sum to $0$.






                share|cite|improve this answer




















                  Your Answer




                  StackExchange.ifUsing("editor", function ()
                  return StackExchange.using("mathjaxEditing", function ()
                  StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix)
                  StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
                  );
                  );
                  , "mathjax-editing");

                  StackExchange.ready(function()
                  var channelOptions =
                  tags: "".split(" "),
                  id: "69"
                  ;
                  initTagRenderer("".split(" "), "".split(" "), channelOptions);

                  StackExchange.using("externalEditor", function()
                  // Have to fire editor after snippets, if snippets enabled
                  if (StackExchange.settings.snippets.snippetsEnabled)
                  StackExchange.using("snippets", function()
                  createEditor();
                  );

                  else
                  createEditor();

                  );

                  function createEditor()
                  StackExchange.prepareEditor(
                  heartbeatType: 'answer',
                  convertImagesToLinks: true,
                  noModals: false,
                  showLowRepImageUploadWarning: true,
                  reputationToPostImages: 10,
                  bindNavPrevention: true,
                  postfix: "",
                  noCode: true, onDemand: true,
                  discardSelector: ".discard-answer"
                  ,immediatelyShowMarkdownHelp:true
                  );



                  );













                   

                  draft saved


                  draft discarded


















                  StackExchange.ready(
                  function ()
                  StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f226023%2fcongruence-modulo-p%23new-answer', 'question_page');

                  );

                  Post as a guest






























                  5 Answers
                  5






                  active

                  oldest

                  votes








                  5 Answers
                  5






                  active

                  oldest

                  votes









                  active

                  oldest

                  votes






                  active

                  oldest

                  votes








                  up vote
                  6
                  down vote



                  accepted










                  We might as well sum to $p-1$. Let $g$ be a primitive root of $p$. Then $1$, $2$, $3$, and so on up to $p-1$ are congruent, in some order, to $g^0$, $g^1$, up to $g^p-2$.
                  So modulo $p$ the terms in our sum are congruent to
                  $g^0$, $g^n$, $g^2n$, up to $g^(p-2)n$. Sum this geometric progression. We get $dfracg^(p-1)n-1g^n-1$. The numerator is congruent to $0$. There is no problem with the denominator, since $nlt p-1$.



                  If one prefers not to use fractions, equivalently note that
                  $$left(1+g^n+g^2n+cdots+g^(p-2)nright)(1-g^n)=1-g^(p-1)n.$$
                  The right-hand side is congruent to $0$. But $1-g^n$ is not congruent to $0$, and the result follows.






                  share|cite|improve this answer


























                    up vote
                    6
                    down vote



                    accepted










                    We might as well sum to $p-1$. Let $g$ be a primitive root of $p$. Then $1$, $2$, $3$, and so on up to $p-1$ are congruent, in some order, to $g^0$, $g^1$, up to $g^p-2$.
                    So modulo $p$ the terms in our sum are congruent to
                    $g^0$, $g^n$, $g^2n$, up to $g^(p-2)n$. Sum this geometric progression. We get $dfracg^(p-1)n-1g^n-1$. The numerator is congruent to $0$. There is no problem with the denominator, since $nlt p-1$.



                    If one prefers not to use fractions, equivalently note that
                    $$left(1+g^n+g^2n+cdots+g^(p-2)nright)(1-g^n)=1-g^(p-1)n.$$
                    The right-hand side is congruent to $0$. But $1-g^n$ is not congruent to $0$, and the result follows.






                    share|cite|improve this answer
























                      up vote
                      6
                      down vote



                      accepted







                      up vote
                      6
                      down vote



                      accepted






                      We might as well sum to $p-1$. Let $g$ be a primitive root of $p$. Then $1$, $2$, $3$, and so on up to $p-1$ are congruent, in some order, to $g^0$, $g^1$, up to $g^p-2$.
                      So modulo $p$ the terms in our sum are congruent to
                      $g^0$, $g^n$, $g^2n$, up to $g^(p-2)n$. Sum this geometric progression. We get $dfracg^(p-1)n-1g^n-1$. The numerator is congruent to $0$. There is no problem with the denominator, since $nlt p-1$.



                      If one prefers not to use fractions, equivalently note that
                      $$left(1+g^n+g^2n+cdots+g^(p-2)nright)(1-g^n)=1-g^(p-1)n.$$
                      The right-hand side is congruent to $0$. But $1-g^n$ is not congruent to $0$, and the result follows.






                      share|cite|improve this answer














                      We might as well sum to $p-1$. Let $g$ be a primitive root of $p$. Then $1$, $2$, $3$, and so on up to $p-1$ are congruent, in some order, to $g^0$, $g^1$, up to $g^p-2$.
                      So modulo $p$ the terms in our sum are congruent to
                      $g^0$, $g^n$, $g^2n$, up to $g^(p-2)n$. Sum this geometric progression. We get $dfracg^(p-1)n-1g^n-1$. The numerator is congruent to $0$. There is no problem with the denominator, since $nlt p-1$.



                      If one prefers not to use fractions, equivalently note that
                      $$left(1+g^n+g^2n+cdots+g^(p-2)nright)(1-g^n)=1-g^(p-1)n.$$
                      The right-hand side is congruent to $0$. But $1-g^n$ is not congruent to $0$, and the result follows.







                      share|cite|improve this answer














                      share|cite|improve this answer



                      share|cite|improve this answer








                      edited Oct 31 '12 at 13:28

























                      answered Oct 31 '12 at 13:13









                      André Nicolas

                      447k36415793




                      447k36415793




















                          up vote
                          6
                          down vote













                          Since the sets of reduced residue classes $1, 2, . . . , (p − 1)$ and $g, 2g, . . . , (p − 1)g$ are the same if $(g,p)=1$ the reason being:
                          if $r_1gequiv r_2gpmod p,$ where $1le r_1< r_2le p-1$
                          $implies r_1equiv r_2$ which is impossible, so $r_1g≢r_2gpmod p$.



                          So, $sum_1le xle p-1 x^n equiv sum_1le yle p-1 (gy)^n pmod p$



                          $implies pmid(g^n-1)sum_1le xle p-1 x^n $



                          If we take $g$ to be a primitive root, $pmid (g^n-1)iff phi(p)=(p-1)mid n$



                          Since $1le n<p-1,p$ can not divide $(g^n-1)implies sum_1le xle p-1 x^nequiv 0pmod p$



                          If $(p-1)mid n, sum_1le xle p-1 x^nequiv sum_1le xle p-11pmod p=p-1equiv -1pmod p $






                          share|cite|improve this answer


























                            up vote
                            6
                            down vote













                            Since the sets of reduced residue classes $1, 2, . . . , (p − 1)$ and $g, 2g, . . . , (p − 1)g$ are the same if $(g,p)=1$ the reason being:
                            if $r_1gequiv r_2gpmod p,$ where $1le r_1< r_2le p-1$
                            $implies r_1equiv r_2$ which is impossible, so $r_1g≢r_2gpmod p$.



                            So, $sum_1le xle p-1 x^n equiv sum_1le yle p-1 (gy)^n pmod p$



                            $implies pmid(g^n-1)sum_1le xle p-1 x^n $



                            If we take $g$ to be a primitive root, $pmid (g^n-1)iff phi(p)=(p-1)mid n$



                            Since $1le n<p-1,p$ can not divide $(g^n-1)implies sum_1le xle p-1 x^nequiv 0pmod p$



                            If $(p-1)mid n, sum_1le xle p-1 x^nequiv sum_1le xle p-11pmod p=p-1equiv -1pmod p $






                            share|cite|improve this answer
























                              up vote
                              6
                              down vote










                              up vote
                              6
                              down vote









                              Since the sets of reduced residue classes $1, 2, . . . , (p − 1)$ and $g, 2g, . . . , (p − 1)g$ are the same if $(g,p)=1$ the reason being:
                              if $r_1gequiv r_2gpmod p,$ where $1le r_1< r_2le p-1$
                              $implies r_1equiv r_2$ which is impossible, so $r_1g≢r_2gpmod p$.



                              So, $sum_1le xle p-1 x^n equiv sum_1le yle p-1 (gy)^n pmod p$



                              $implies pmid(g^n-1)sum_1le xle p-1 x^n $



                              If we take $g$ to be a primitive root, $pmid (g^n-1)iff phi(p)=(p-1)mid n$



                              Since $1le n<p-1,p$ can not divide $(g^n-1)implies sum_1le xle p-1 x^nequiv 0pmod p$



                              If $(p-1)mid n, sum_1le xle p-1 x^nequiv sum_1le xle p-11pmod p=p-1equiv -1pmod p $






                              share|cite|improve this answer














                              Since the sets of reduced residue classes $1, 2, . . . , (p − 1)$ and $g, 2g, . . . , (p − 1)g$ are the same if $(g,p)=1$ the reason being:
                              if $r_1gequiv r_2gpmod p,$ where $1le r_1< r_2le p-1$
                              $implies r_1equiv r_2$ which is impossible, so $r_1g≢r_2gpmod p$.



                              So, $sum_1le xle p-1 x^n equiv sum_1le yle p-1 (gy)^n pmod p$



                              $implies pmid(g^n-1)sum_1le xle p-1 x^n $



                              If we take $g$ to be a primitive root, $pmid (g^n-1)iff phi(p)=(p-1)mid n$



                              Since $1le n<p-1,p$ can not divide $(g^n-1)implies sum_1le xle p-1 x^nequiv 0pmod p$



                              If $(p-1)mid n, sum_1le xle p-1 x^nequiv sum_1le xle p-11pmod p=p-1equiv -1pmod p $







                              share|cite|improve this answer














                              share|cite|improve this answer



                              share|cite|improve this answer








                              edited Oct 31 '12 at 15:41

























                              answered Oct 31 '12 at 13:17









                              lab bhattacharjee

                              216k14153266




                              216k14153266




















                                  up vote
                                  2
                                  down vote













                                  Call
                                  $$ p_n = sum_k=1^p k^n. $$
                                  Since, by Fermat's little theorem, $p_nequiv p_n-p+1pmodp$ for all $n > p-1$, it is sufficient to show that
                                  $$ forall nin[0,p-2],quad p_nequiv 0pmodp, $$
                                  where the cases $n=0$ and $n=1$ are trivial. Since:
                                  $$ x(x-1)cdotldotscdot(x-p+1)equiv x^p-xpmodp, $$
                                  if $e_k(x_1,ldots,x_p)$ is the $k$-th elementary symmetric polynomial, we have:
                                  $$ forall kin[0,p-2],quad e_k(1,ldots,p)equiv 0pmodp, $$
                                  so, in virtue of Newton-Girard identities and induction we have:
                                  $$forall kin[0,p-2],quad p_kequiv 0pmodp, $$
                                  QED.






                                  share|cite|improve this answer


























                                    up vote
                                    2
                                    down vote













                                    Call
                                    $$ p_n = sum_k=1^p k^n. $$
                                    Since, by Fermat's little theorem, $p_nequiv p_n-p+1pmodp$ for all $n > p-1$, it is sufficient to show that
                                    $$ forall nin[0,p-2],quad p_nequiv 0pmodp, $$
                                    where the cases $n=0$ and $n=1$ are trivial. Since:
                                    $$ x(x-1)cdotldotscdot(x-p+1)equiv x^p-xpmodp, $$
                                    if $e_k(x_1,ldots,x_p)$ is the $k$-th elementary symmetric polynomial, we have:
                                    $$ forall kin[0,p-2],quad e_k(1,ldots,p)equiv 0pmodp, $$
                                    so, in virtue of Newton-Girard identities and induction we have:
                                    $$forall kin[0,p-2],quad p_kequiv 0pmodp, $$
                                    QED.






                                    share|cite|improve this answer
























                                      up vote
                                      2
                                      down vote










                                      up vote
                                      2
                                      down vote









                                      Call
                                      $$ p_n = sum_k=1^p k^n. $$
                                      Since, by Fermat's little theorem, $p_nequiv p_n-p+1pmodp$ for all $n > p-1$, it is sufficient to show that
                                      $$ forall nin[0,p-2],quad p_nequiv 0pmodp, $$
                                      where the cases $n=0$ and $n=1$ are trivial. Since:
                                      $$ x(x-1)cdotldotscdot(x-p+1)equiv x^p-xpmodp, $$
                                      if $e_k(x_1,ldots,x_p)$ is the $k$-th elementary symmetric polynomial, we have:
                                      $$ forall kin[0,p-2],quad e_k(1,ldots,p)equiv 0pmodp, $$
                                      so, in virtue of Newton-Girard identities and induction we have:
                                      $$forall kin[0,p-2],quad p_kequiv 0pmodp, $$
                                      QED.






                                      share|cite|improve this answer














                                      Call
                                      $$ p_n = sum_k=1^p k^n. $$
                                      Since, by Fermat's little theorem, $p_nequiv p_n-p+1pmodp$ for all $n > p-1$, it is sufficient to show that
                                      $$ forall nin[0,p-2],quad p_nequiv 0pmodp, $$
                                      where the cases $n=0$ and $n=1$ are trivial. Since:
                                      $$ x(x-1)cdotldotscdot(x-p+1)equiv x^p-xpmodp, $$
                                      if $e_k(x_1,ldots,x_p)$ is the $k$-th elementary symmetric polynomial, we have:
                                      $$ forall kin[0,p-2],quad e_k(1,ldots,p)equiv 0pmodp, $$
                                      so, in virtue of Newton-Girard identities and induction we have:
                                      $$forall kin[0,p-2],quad p_kequiv 0pmodp, $$
                                      QED.







                                      share|cite|improve this answer














                                      share|cite|improve this answer



                                      share|cite|improve this answer








                                      edited Sep 10 at 13:54









                                      darij grinberg

                                      9,47132960




                                      9,47132960










                                      answered Oct 31 '12 at 13:18









                                      Jack D'Aurizio♦

                                      275k32268641




                                      275k32268641




















                                          up vote
                                          0
                                          down vote













                                          EDIT:This is a wrong answer($r^n equiv t^n pmod p not Rightarrow r equiv t pmod p$). See the comments.



                                          For $n=1$ is correct.



                                          Now for any $n in 1,2,ldots,p-2$ we have that $1,2,ldots,p-1equiv 1,2^n,ldots,(p-1)^n pmod p$ since $r^n equiv t^n pmod p iff r equiv t pmod p$. Hence $$sum_t=1^pt^n equiv sum_t=1^pt equiv 0 pmodp .$$






                                          share|cite|improve this answer






















                                          • I don't thing the $equiv$ is true. For example, if $n=2$, the sequence $1,2^2,3^2,...,(p-1)^2$ only has $fracp-12$ distinct elements
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:52






                                          • 1




                                            Sorry, but this isn't completely watertight. With $n = 2$ and $p = 5$, you get $1^2, 2^2, 3^2, 4^2 cong 1, 4, 4, 1$
                                            – Arthur
                                            Oct 31 '12 at 12:53










                                          • Note that $1^2equiv (p-1)^2pmod p$, but $1notequiv p-1pmod p$
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:53














                                          up vote
                                          0
                                          down vote













                                          EDIT:This is a wrong answer($r^n equiv t^n pmod p not Rightarrow r equiv t pmod p$). See the comments.



                                          For $n=1$ is correct.



                                          Now for any $n in 1,2,ldots,p-2$ we have that $1,2,ldots,p-1equiv 1,2^n,ldots,(p-1)^n pmod p$ since $r^n equiv t^n pmod p iff r equiv t pmod p$. Hence $$sum_t=1^pt^n equiv sum_t=1^pt equiv 0 pmodp .$$






                                          share|cite|improve this answer






















                                          • I don't thing the $equiv$ is true. For example, if $n=2$, the sequence $1,2^2,3^2,...,(p-1)^2$ only has $fracp-12$ distinct elements
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:52






                                          • 1




                                            Sorry, but this isn't completely watertight. With $n = 2$ and $p = 5$, you get $1^2, 2^2, 3^2, 4^2 cong 1, 4, 4, 1$
                                            – Arthur
                                            Oct 31 '12 at 12:53










                                          • Note that $1^2equiv (p-1)^2pmod p$, but $1notequiv p-1pmod p$
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:53












                                          up vote
                                          0
                                          down vote










                                          up vote
                                          0
                                          down vote









                                          EDIT:This is a wrong answer($r^n equiv t^n pmod p not Rightarrow r equiv t pmod p$). See the comments.



                                          For $n=1$ is correct.



                                          Now for any $n in 1,2,ldots,p-2$ we have that $1,2,ldots,p-1equiv 1,2^n,ldots,(p-1)^n pmod p$ since $r^n equiv t^n pmod p iff r equiv t pmod p$. Hence $$sum_t=1^pt^n equiv sum_t=1^pt equiv 0 pmodp .$$






                                          share|cite|improve this answer














                                          EDIT:This is a wrong answer($r^n equiv t^n pmod p not Rightarrow r equiv t pmod p$). See the comments.



                                          For $n=1$ is correct.



                                          Now for any $n in 1,2,ldots,p-2$ we have that $1,2,ldots,p-1equiv 1,2^n,ldots,(p-1)^n pmod p$ since $r^n equiv t^n pmod p iff r equiv t pmod p$. Hence $$sum_t=1^pt^n equiv sum_t=1^pt equiv 0 pmodp .$$







                                          share|cite|improve this answer














                                          share|cite|improve this answer



                                          share|cite|improve this answer








                                          edited Oct 31 '12 at 13:12

























                                          answered Oct 31 '12 at 12:50









                                          P..

                                          13.2k22346




                                          13.2k22346











                                          • I don't thing the $equiv$ is true. For example, if $n=2$, the sequence $1,2^2,3^2,...,(p-1)^2$ only has $fracp-12$ distinct elements
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:52






                                          • 1




                                            Sorry, but this isn't completely watertight. With $n = 2$ and $p = 5$, you get $1^2, 2^2, 3^2, 4^2 cong 1, 4, 4, 1$
                                            – Arthur
                                            Oct 31 '12 at 12:53










                                          • Note that $1^2equiv (p-1)^2pmod p$, but $1notequiv p-1pmod p$
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:53
















                                          • I don't thing the $equiv$ is true. For example, if $n=2$, the sequence $1,2^2,3^2,...,(p-1)^2$ only has $fracp-12$ distinct elements
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:52






                                          • 1




                                            Sorry, but this isn't completely watertight. With $n = 2$ and $p = 5$, you get $1^2, 2^2, 3^2, 4^2 cong 1, 4, 4, 1$
                                            – Arthur
                                            Oct 31 '12 at 12:53










                                          • Note that $1^2equiv (p-1)^2pmod p$, but $1notequiv p-1pmod p$
                                            – Thomas Andrews
                                            Oct 31 '12 at 12:53















                                          I don't thing the $equiv$ is true. For example, if $n=2$, the sequence $1,2^2,3^2,...,(p-1)^2$ only has $fracp-12$ distinct elements
                                          – Thomas Andrews
                                          Oct 31 '12 at 12:52




                                          I don't thing the $equiv$ is true. For example, if $n=2$, the sequence $1,2^2,3^2,...,(p-1)^2$ only has $fracp-12$ distinct elements
                                          – Thomas Andrews
                                          Oct 31 '12 at 12:52




                                          1




                                          1




                                          Sorry, but this isn't completely watertight. With $n = 2$ and $p = 5$, you get $1^2, 2^2, 3^2, 4^2 cong 1, 4, 4, 1$
                                          – Arthur
                                          Oct 31 '12 at 12:53




                                          Sorry, but this isn't completely watertight. With $n = 2$ and $p = 5$, you get $1^2, 2^2, 3^2, 4^2 cong 1, 4, 4, 1$
                                          – Arthur
                                          Oct 31 '12 at 12:53












                                          Note that $1^2equiv (p-1)^2pmod p$, but $1notequiv p-1pmod p$
                                          – Thomas Andrews
                                          Oct 31 '12 at 12:53




                                          Note that $1^2equiv (p-1)^2pmod p$, but $1notequiv p-1pmod p$
                                          – Thomas Andrews
                                          Oct 31 '12 at 12:53










                                          up vote
                                          0
                                          down vote













                                          WLOG assume $nk = p-1$ (if $n$ had factors coprime to $p-1$ they would just permute the sum, not change it's value, so remove them).



                                          Put $S = (mathbb Z/pmathbb Z)^times$ and let $S^n$ denote the subgroup $ s in S $. The kernel of the map $S to S^n$ is the set of solutions of $z^n=1$ so $n|S^n| = |S|$. Details about subgroup here



                                          This tells us that (mod p) the sum is equal to $n$ times the sum of the elements of $S^n$ so we just need to show that $$sum_s in S^ns equiv 0 pmod p$$ but since this is the set of solutions of $z^n-1=0$ by Vieta's formula they sum to $0$.






                                          share|cite|improve this answer
























                                            up vote
                                            0
                                            down vote













                                            WLOG assume $nk = p-1$ (if $n$ had factors coprime to $p-1$ they would just permute the sum, not change it's value, so remove them).



                                            Put $S = (mathbb Z/pmathbb Z)^times$ and let $S^n$ denote the subgroup $ s in S $. The kernel of the map $S to S^n$ is the set of solutions of $z^n=1$ so $n|S^n| = |S|$. Details about subgroup here



                                            This tells us that (mod p) the sum is equal to $n$ times the sum of the elements of $S^n$ so we just need to show that $$sum_s in S^ns equiv 0 pmod p$$ but since this is the set of solutions of $z^n-1=0$ by Vieta's formula they sum to $0$.






                                            share|cite|improve this answer






















                                              up vote
                                              0
                                              down vote










                                              up vote
                                              0
                                              down vote









                                              WLOG assume $nk = p-1$ (if $n$ had factors coprime to $p-1$ they would just permute the sum, not change it's value, so remove them).



                                              Put $S = (mathbb Z/pmathbb Z)^times$ and let $S^n$ denote the subgroup $ s in S $. The kernel of the map $S to S^n$ is the set of solutions of $z^n=1$ so $n|S^n| = |S|$. Details about subgroup here



                                              This tells us that (mod p) the sum is equal to $n$ times the sum of the elements of $S^n$ so we just need to show that $$sum_s in S^ns equiv 0 pmod p$$ but since this is the set of solutions of $z^n-1=0$ by Vieta's formula they sum to $0$.






                                              share|cite|improve this answer












                                              WLOG assume $nk = p-1$ (if $n$ had factors coprime to $p-1$ they would just permute the sum, not change it's value, so remove them).



                                              Put $S = (mathbb Z/pmathbb Z)^times$ and let $S^n$ denote the subgroup $ s in S $. The kernel of the map $S to S^n$ is the set of solutions of $z^n=1$ so $n|S^n| = |S|$. Details about subgroup here



                                              This tells us that (mod p) the sum is equal to $n$ times the sum of the elements of $S^n$ so we just need to show that $$sum_s in S^ns equiv 0 pmod p$$ but since this is the set of solutions of $z^n-1=0$ by Vieta's formula they sum to $0$.







                                              share|cite|improve this answer












                                              share|cite|improve this answer



                                              share|cite|improve this answer










                                              answered Oct 31 '12 at 13:59









                                              sperners lemma

                                              1,415929




                                              1,415929



























                                                   

                                                  draft saved


                                                  draft discarded















































                                                   


                                                  draft saved


                                                  draft discarded














                                                  StackExchange.ready(
                                                  function ()
                                                  StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f226023%2fcongruence-modulo-p%23new-answer', 'question_page');

                                                  );

                                                  Post as a guest













































































                                                  這個網誌中的熱門文章

                                                  tkz-euclide: tkzDrawCircle[R] not working

                                                  How to combine Bézier curves to a surface?

                                                  1st Magritte Awards