Sharedwww / primes.dviOpen in CoCalc
����;� TeX output 1998.11.29:0439������ufv�������6fv����szt�D��tG�G�cmr17�Arithmetic�7tProgressions�of�Primes��lύ������Q|�X�Qcmr12�W.��Stein�����2��K�cmsy8����������ύ��U�No��rv�em�b�S�er��29,�1998��*�Y������-�"V

cmbx10�Abstract��
_G���K�y

cmr10�Let�#�2�
�b>

cmmi10�n��b�Ge�an�ev���en�in�teger.�a=What�is�the�longest�arithmetic�progres-����Qsion��ɵp;�UPp�kɲ+�2�n;��:���:�:��UG;�p��+�2�na��ɲwith��p��prime?�O�In�this�pap�Ger�w���e�guess�some����Qansw���ers.��)=��6�3�"V
�3
cmbx10�W���\arning:�k�K�y
�3
cmr10�The�mfollo��!wing�is�just�for�fun!�
1�There�could�b�M�e�a�w�ell-kno�wn��
����6conjecture��fwhic��!h�sup�M�ercedes�all�of�this.��!����6�5��N�ffcmbx12�1��NL�In���tro�s3duction��q���6�It��fis�easy�to�sho��!w�(see�Prop�M�osition�3.1)�that�if���ߍ��/��6�b>
�3
cmmi10�p;��bp�n�+�2�;�p��+�4����6are��fall�prime,�then��p�
��=�3.���Similarly��e,��fif������)�p;��bp�n�+�6�;�p��+�12�;�p��+�18�;�p��+�24����6are�,�all�prime,�E5then��p�
��=�5�,�and�the�progression�is�th��!us�5�;��b�11�;��17�;��23�;��29�:�,�There����6are��fprobably�innitely�man��!y�examples�where������1�p;��bp�n�+�6�;�p��+�12�;�p��+�18�;����6�are��8all�prime,�.�but�one�w��!ould�not�dare�think�of�pro�ving�this�here�as�it�is����6probably��fmore�dicult�than�the�t��!win�primes�conjecture.����GF��eor���eac��!h�ev�en�in�teger�2�n�,�O�let��M�1��(2�n�)�to�b�M�e�the�maximal�length�of�an����6arithmetic���progression�of�primes�with�gap�2�n�.�K�Th��!us��M�1��(2)��<=�3,��M��(6)�=����65.�õLet�HY�M���z�1��	�(2�n�)�b�M�e�the�maxim��!um�length�of�arithmetic�progressions�of�gap����62�n���whic��!h�o�M�ccur�innitely�often.�	,Th�us�one�conjectures�that��M���z�1��	�(2)�g�=�2,����6�M���z�1��	�(6)�
�=�4.����GIs�ݏthere�an�example�in�whic��!h��M���z�1��	�(2�n�)�f��<�M�1��(2�n�)����7!",�
�3
cmsy10���1?�ݏIs�there�an�upp�M�er����6b�M�ound��fon�the�p�ossible�v��dDalues��M�1��(2�n�)?��6���ff��p�
L͍���Q
��-=�q�%cmsy6�����a�&o���		cmr9�UC�TBerk��9eley��:�,�Departmen�t�of�Mathematics,��,ߤN		cmtt9�[email protected]������C3�1����*�ufv�������6fv���홊���6�2��NL�T���fable��q���6�The�k��M��Ȯ�|{Ycmr8�?���3�column�w��!as�computed�b�y�searc�hing�the�rst�100000�or�so�primes��
����6for�"�an�arithmetic�progression�of�primes,�A�ha��!ving�gap�2�n�.�R�The��M��Ȯ�1�?��&>�column����6w��!as�2�computed�b�y�doing�the�same�searc�h,�I�but�requiring�that�at�least��;�':
�3
cmti10�two��dis-����6tinct��progressions�w��!ere�found�(usually�there�w�ere�man�y�more).�=qBy�Prop�M�o-����6sition�ڪ3.1,�'�the�computed�v��dDalues�of��M��Ȯ�?��	^<�are�the�actual�v�alue�of��M�1��,�'�except����6p�M�ossibly��ffor�the�last�four�ro��!ws�in�the�table.���wl����Q����/�ff0d
�	�����ͤ}�
��ff�3����2�n���
��ff���_��2�n�2�A�}�
��ff�����M��Ȯ�?��
P_�}�
��ff����/��M��Ȯ�1�?���g�}�
��ff�����ݹExamples�9��}�
��ff���z�ff0d
��ff0d
�����ͤ}�
��ff�|Ο��2���
��ff���c6Z2�6)��}�
��ff�����n3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff���
������ͤ}�
��ff�|Ο��4���
��ff����X2�����2��8���}�
��ff�����n�3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�|Ο��6���
��ff���\�L2�n����3�/x}�}�
��ff�����n5�
���}�
��ff����D�4�፟}�
��ff������5�F�c�}�
��ff�������ͤ}�
��ff�|Ο��8���
��ff����X2�����3��8���}�
��ff�����n�3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����10���
��ff���\�L2�n����5�/x}�}�
��ff�����n3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff����ff0d
�	�����ͤ}�
��ff�����12���
��ff���Z%J2�����2��.���n�3�-{�}�
��ff�����n5�
���}�
��ff����D�4�፟}�
��ff������5�F�c�}�
��ff�������ͤ}�
��ff�����14���
��ff���\�L2�n����7�/x}�}�
��ff�����n3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����16���
��ff����X2�����4��8���}�
��ff�����n�2�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����18���
��ff���Z%J2�n����3�����2��1��}�
��ff�����n�4�
���}�
��ff����D�4�፟}�
��ff������5�F�c�}�
��ff�������ͤ}�
��ff�����20���
��ff���Z%J2�����2��.���n�5�-{�}�
��ff�����n3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff����ff0d
�	�����ͤ}�
��ff�����22���
��ff���Y�2�n����11�,���}�
��ff�����n2�
���}�
��ff����D�2�፟}�
��ff������7�F�c�}�
��ff�������ͤ}�
��ff�����24���
��ff���Z%J2�����3��.���n�3�-{�}�
��ff�����n4�
���}�
��ff����D�4�፟}�
��ff������59�Ajɟ}�
��ff�������ͤ}�
��ff�����26���
��ff���Y�2�n����13�,���}�
��ff�����n2�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����28���
��ff���Z%J2�����2��.���n�7�-{�}�
��ff�����n2�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����30���
��ff���U�=2�n����3����5�(�n�}�
��ff�����n6�
���}�
��ff����D�6�፟}�
��ff������7,��f107�/Ƅ�}�
��ff����ff0d
�	�����ͤ}�
��ff�����32���
��ff����X2�����5��8���}�
��ff�����n�2�
���}�
��ff����D�2�፟}�
��ff������5�F�c�}�
��ff�������ͤ}�
��ff�����34���
��ff���Y�2�n����17�,���}�
��ff�����n3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����36���
��ff���W�H2�����2��.���n�3�����2��/x}�}�
��ff�����n�4�
���}�
��ff����D�4�፟}�
��ff������31�Ajɟ}�
��ff�������ͤ}�
��ff�����38���
��ff���Y�2�n����19�,���}�
��ff�����n3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����40���
��ff���Z%J2�����3��.���n�5�-{�}�
��ff�����n3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff����ff0d
�	�����ͤ}�
��ff�����42���
��ff���U�=2�n����3����7�(�n�}�
��ff�����n5�
���}�
��ff����D�4�፟}�
��ff������5�F�c�}�
��ff�������ͤ}�
��ff�����44���
��ff���Wh}2�����2��.���n�11�*[��}�
��ff�����n2�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff�������ͤ}�
��ff�����46���
��ff���Y�2�n����23�,���}�
��ff�����n2�
���}�
��ff����D�2�፟}�
��ff������7�F�c�}�
��ff�������ͤ}�
��ff�����48���
��ff���Z%J2�����4��.���n�3�-{�}�
��ff�����n5�
���}�
��ff����D�4�፟}�
��ff������5�F�c�}�
��ff�������ͤ}�
��ff�����50���
��ff���Z%J2�n����5�����2��1��}�
��ff�����n�3�
���}�
��ff����D�2�፟}�
��ff������3�F�c�}�
��ff����ff0d
�	�����ͤ}�
��ff�4���210���
��ff���O#/2�n����3����5����7�"�}�
��ff����$�10�伟}� ��ff����D�9�፟}� ��ff������199�;�/�}� ��ff�������ͤ}� ��ff� Fg���2310��� ��ff���E�S2�n����3����5����7����11����}� ��ff�����n9� ���}� ��ff����D�9�፟}� ��ff������3823�6w��}� ��ff�������ͤ}� ��ff������30030��� ��ff���<Gx2�n����3����5����7����11����13�:��}� ��ff����$�12�伟}�
��ff���Ƈ�12�$��}� ��ff������23143,��M1498141��͟}� ��ff�������ͤ}� ��ff��͟��510510��� ��ff���2ٜ2�n����3����5����7����11����13����17��͟}� ��ff����$�13�伟}�
��ff����kq?�{�}�
��ff������766439�+�a�}�
��ff����ff0d
��������C32����	�ufv�������6fv���홊���6�3��NL�Conclusions��q�����6�Prop�Y�osition�2�3.1.����$��L��p�et��^�q�0��b�e�a�prime.���If��q�o:�:��� �3 msbm10�-� ��2�n��then��M�1��(2�n�)����q�0��and��M���z�1�� �(2�n�)���� ����6�q�&����/�1�.���F��)urthermor��p�e,��Y�M�1��(2�n�)� �=��q��i ��t�q��is�involve��p�d�in�arithmetic�pr�o�gr�ession�of����6length����q�O}�with�gap��2�n�.��������6Pr��p�o�of.���XON�Reduce��fthe�arithmetic�progression�������)U�p;��bp�n�+�2�n;��:��1:�:��';�p��+�2�na����6�mo�M�dulo�=Z�q�d��.� ��Since�2�n��is�a�unit�mo�dulo��q�d��,��if�the�length�of�the�arithmetic�� ����6progression���is������q�#H�then�one�of�the�terms�in�the�arithmetic�progression�is����6divisible��fb��!y��q�d��,�hence�equal�to��q��.����萄d,ff���:�ff���Ɖff����d,ff������GIn��the�table�ab�M�o��!v�e,�Јthere��is�not�a�single�example�in�whic��!h��q�o:�-� ��2�n�,��M�1��(2�n�)�=����6�q� ��and��f�q��is�not�the� rst�prime�in�the�arithmetic.���This�suggests������6�Question�2�3.2.����$6�L��p�et����q�o:�-�
��2�n��b�e�a�prime�and�supp�ose��M�1��(2�n�)�
�=��q�d��.�	vThen���must���������q�d�;��bq���+�n�2�n;��:��1:�:��';�q��+�n�2�n�(�q����1)����6�al��Fl���b��p�e�prime?����G�The��fdata�also�suggests�the�follo��!wing������6�Conjecture�2�3.3.�������Both��f�M�1��(2�n�)�and��M���z�1��	�(2�n�)�are���
��2�for�all��n�.����GThe��
conjecture�that��M�1��(2�n�)�ua���2��
is�similar�to�Golbac��!h's�conjecture.�j�It��
����6asserts��<that�an��!y�ev�en�in�teger�is�the�dierence�of�t�w�o�primes;���the�conjecture����6that��f�M���z�1��	�(2�n�)�
����2�asserts�that�this�can�b�M�e�done�in�innitely�man��!y�w�a�ys.����GIt��fseems�conceiv��dDable�that��M�1��(2�n�!)�is�un��!b�M�ounded�as�a�function�of��n�.������C33�������;�ufv�	�;�':
�3
cmti10�:���
�3
msbm10�7!",�
�3
cmsy10�6�b>
�3
cmmi10�5��N�ffcmbx12�3�"V
�3
cmbx10�-�"V

cmbx10�,ߤN		cmtt9�&o���		cmr9�q�%cmsy6��K�cmsy8�|{Ycmr8�X�Qcmr12�D��tG�G�cmr17�K�y
�3
cmr10�
�b>

cmmi10�K�y

cmr10�&x������`