<?xml version="1.0" encoding="UTF-8"?><!DOCTYPE article  PUBLIC "-//NLM//DTD Journal Publishing DTD v3.0 20080202//EN" "http://dtd.nlm.nih.gov/publishing/3.0/journalpublishing3.dtd"><article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" dtd-version="3.0" xml:lang="en" article-type="research article"><front><journal-meta><journal-id journal-id-type="publisher-id">AM</journal-id><journal-title-group><journal-title>Applied Mathematics</journal-title></journal-title-group><issn pub-type="epub">2152-7385</issn><publisher><publisher-name>Scientific Research Publishing</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.4236/am.2014.513183</article-id><article-id pub-id-type="publisher-id">AM-47626</article-id><article-categories><subj-group subj-group-type="heading"><subject>Articles</subject></subj-group><subj-group subj-group-type="Discipline-v2"><subject>COMPUTER SCIENCE &amp; COMMUNICATIONS</subject><subject>ENGINEERING</subject><subject>PHYSICS &amp; MATHEMATICS</subject></subj-group></article-categories><title-group><article-title>Block Matrix Representation of a Graph Manifold Linking Matrix Using Continued Fractions</article-title></title-group><contrib-group><contrib contrib-type="author" xlink:type="simple"><name name-style="western"><surname>Fernando</surname><given-names>I. Becerra López</given-names></name><xref ref-type="aff" rid="aff1"><sup>1</sup></xref><xref ref-type="corresp" rid="cor1"><sup>*</sup></xref></contrib><contrib contrib-type="author" xlink:type="simple"><name name-style="western"><surname>Vladimir</surname><given-names>N. Efremov</given-names></name><xref ref-type="aff" rid="aff1"><sup>1</sup></xref><xref ref-type="corresp" rid="cor1"><sup>*</sup></xref></contrib><contrib contrib-type="author" xlink:type="simple"><name name-style="western"><surname>Alfonso</surname><given-names>M. Hernandez Magdaleno</given-names></name><xref ref-type="aff" rid="aff1"><sup>1</sup></xref><xref ref-type="corresp" rid="cor1"><sup>*</sup></xref></contrib></contrib-group><aff id="aff1"><addr-line>Mathematics Department, CUCEI, University of Guadalajara, Guadalajara, Mexico</addr-line></aff><author-notes><corresp id="cor1">* E-mail:<email>mail@ferdx.com(FIBL)</email>;<email>efremov@cencar.udg.mx(VNE)</email>;<email>137mag@gmail.com(AMHM)</email>;</corresp></author-notes><pub-date pub-type="epub"><day>07</day><month>07</month><year>2014</year></pub-date><volume>05</volume><issue>13</issue><fpage>1894</fpage><lpage>1902</lpage><history><date date-type="received"><day>19</day>	<month>April</month>	<year>2014</year></date><date date-type="rev-recd"><day>27</day>	<month>May</month>	<year>2014</year>	</date><date date-type="accepted"><day>6</day>	<month>June</month>	<year>2014</year></date></history><permissions><copyright-statement>&#169; Copyright  2014 by authors and Scientific Research Publishing Inc. </copyright-statement><copyright-year>2014</copyright-year><license><license-p>This work is licensed under the Creative Commons Attribution International License (CC BY). http://creativecommons.org/licenses/by/4.0/</license-p></license></permissions><abstract><p>We consider the block matrices and 3-dimensional graph manifolds
associated with a special type of tree graphs. We demonstrate that the linking
matrices of these graph manifolds coincide with the reduced matrices obtained
from the Laplacian block matrices by means of Gauss partial diagonalization
procedure described explicitly by W. Neumann. The linking matrix is an
important topological invariant of a graph manifold which is possible to
interpret as a matrix of coupling constants of gauge interaction in
Kaluza-Klein approach, where 3-dimensional graph manifold plays the role of
internal space in topological 7-dimensional BF theory. The Gauss-Neumann method
gives us a simple algorithm to calculate the linking matrices of graph
manifolds and thus the coupling constants matrices.</p></abstract><kwd-group><kwd>Graph Manifolds</kwd><kwd> Continued Fractions</kwd><kwd> Laplacian Matrices</kwd><kwd> Kaluza-Klein</kwd></kwd-group></article-meta></front><body><sec id="s1"><title>1. Introduction</title><p>Graphs can serve as a universal remedy for the codification and classification of topological spaces, matrices, dynamical systems, etc. In this article, we consider the following question: how the topological invariants of manifold corresponding to a tree graph (graph manifold) can be calculated using the method of Gauss-Neumann partial diagonalization of Laplacian matrix defined for the same graph. This calculation can be useful in multidimensional models of Kaluza-Klein type, where coupling constants of gauge interactions are simulated by the rational linking matrices of the internal space [<xref ref-type="bibr" rid="scirp.47626-ref1">1</xref>] . We constructed various models [<xref ref-type="bibr" rid="scirp.47626-ref2">2</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref3">3</xref>] where the role of internal spaces is played by a specific family of 3-dimensional graph manifolds, whose rational linking matrices describe the hierarchy of gauge coupling constants of the real universe. The basic structure blocks of these graph manifolds are Seifert fibered Brieskorn homology spheres, defined as the link of Brieskorn singularity</p><p><inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c49afb35-bfad-4661-8488-292c6722c8af.png" xlink:type="simple"/></inline-formula>, with <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d1a99c0a-da20-439e-b475-df246b31e61c.png" xlink:type="simple"/></inline-formula> pairwise relatively prime positive numbers [<xref ref-type="bibr" rid="scirp.47626-ref4">4</xref>] . Bh-spheres belong to the class of Seifert fibered homology spheres (Sfh-spheres). On each of these manifolds, there exists a Seifert fibration with unnormalized Seifert invariants <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f3d4f4ea-0680-477c-acff-de7a97343825.png" xlink:type="simple"/></inline-formula> subject to  <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0c040cae-be2c-4721-b0df-c787cfa7c432.png" xlink:type="simple"/></inline-formula>, where <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\6067e71c-d9db-45c0-a0e5-ad4675763872.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b2ae2ad2-3f5c-4259-9988-395c8535bd1f.png" xlink:type="simple"/></inline-formula> is its rational Euler number, the well known topological invariant of a Bh-sphere. The topological operation known as plumbing [<xref ref-type="bibr" rid="scirp.47626-ref5">5</xref>] is used to past together Bh-spheres according to plumbing diagrams (<xref ref-type="fig" rid="fig4">Figure 4</xref>) which can be transformed in plumbing graphs (<xref ref-type="fig" rid="fig3">Figure 3</xref>). This type of graphs and their Laplacian block matrices is the main object of attention in this article.</p><p>The paper is organized as follows. In section 2 we define the type of tree graphs considered in this paper (plumbing graphs) and Laplacian matrices for these plumbing graphs. We recall also the Gauss-Neumann method of partial diagonalization by means of which we obtain the reduced rational tridiagonal matrix for each plumbing graph. In section 3 we construct graph manifolds codified by the plumbing graphs defined in section 2 and calculate the main topological invariant for these 3-dimensional manifolds, namely rational linking matrix. Then we demonstrate that the linking matrices of these graph manifolds coincide with the reduced matrices obtained from the Laplacian block matrices by means of Gauss-Neumann partial diagonalization procedure. Finally, we conclude formulating our main results and considering an example of their application for the topological field theory.</p></sec><sec id="s2"><title>2. Block Matrix Representation for a Graph <img src="htmlimages\6-7402261x\75a5eddc-d018-4c3f-b109-bc3198dd7af6.png" width="43.75" height="47.5" /> of Tree Type</title><p>We begin from the definition of graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\801ba1bd-6833-49bf-88a4-88fdf15a964b.png" xlink:type="simple"/></inline-formula> as a finite one-dimensional simplicial complex, which does not contain multiple edges and loops, i.e. we consider only the graph of tree type. An integer weight <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\77ec1935-03b3-4f5f-b583-6c92fb7ee920.png" xlink:type="simple"/></inline-formula> is assigned to each vertex of<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\9aaa6c6b-0394-4031-a574-4df94c816307.png" xlink:type="simple"/></inline-formula>. Vertices with at least three edges are called nodes. For simplicity we shall use graphs with nodes of minimal valence (n = 3) only (a generalization to <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f27092be-b232-471b-aca1-d8c5a8c22e5b.png" xlink:type="simple"/></inline-formula> is obvious). Suppose that the set of nodes <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3575d431-cb76-410c-ab1d-9761dbf305ce.png" xlink:type="simple"/></inline-formula> of the graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\89f8f790-799c-4028-a858-36231ebe9ff9.png" xlink:type="simple"/></inline-formula> is non-empty. Considering the graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c2d8265e-afd3-4567-9a03-aae3b911610e.png" xlink:type="simple"/></inline-formula> as a one-dimensional simplicial complex, we take the complement<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1872d9c2-11e2-4d55-863d-d37a128b46b8.png" xlink:type="simple"/></inline-formula>. This complement is the disjoint union of straight line segments which are the maximal chains of<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1053da7c-c8ba-426f-9058-feecdc61926e.png" xlink:type="simple"/></inline-formula>. <xref ref-type="fig" rid="fig1">Figure 1</xref> shows a maximal internal chain <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\4bc21f72-73b2-4760-a205-83f242d83ea5.png" xlink:type="simple"/></inline-formula> of length k between two nodes <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d9d98740-babf-4315-84af-437a66b4f1c5.png" xlink:type="simple"/></inline-formula> and<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3fd9aba7-f911-482d-803c-b4e9a065b464.png" xlink:type="simple"/></inline-formula>, with weights <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ba7ece3f-491c-46c5-b33e-65cfe973052a.png" xlink:type="simple"/></inline-formula> embedded in a tree graph<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\70d7003c-976d-42e4-a3d5-b3b952cdcca1.png" xlink:type="simple"/></inline-formula>. The chain is maximal because it cannot be included in some larger chain. <xref ref-type="fig" rid="fig2">Figure 2</xref> shows a maximal terminal chain <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d4dd304d-77a8-46e2-a594-4485883d4101.png" xlink:type="simple"/></inline-formula> of length k also with weights<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\923ddae7-eb40-4b67-9947-24a696c7c4db.png" xlink:type="simple"/></inline-formula>.</p><p>In this paper we shall considered only the simplest type of graphs which are called plumbing graphs. An example of plumbing graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a5be935b-affc-4c22-8aed-dab4342a20e8.png" xlink:type="simple"/></inline-formula> is given in <xref ref-type="fig" rid="fig3">Figure 3</xref>. This type of graphs is used to codify the plumbing graph manifolds [<xref ref-type="bibr" rid="scirp.47626-ref5">5</xref>] which will be constructed in the following section, where it will be clear why weights of plumbing graph are called Euler numbers. In <xref ref-type="fig" rid="fig3">Figure 3</xref> the Euler numbers <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\980ef17e-b83b-429c-a5b4-1b8bd175230d.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f85a59ed-0d9e-4e65-8f53-3d28e30c3187.png" xlink:type="simple"/></inline-formula> decorate the vertices with valence 1 and 2. The nodes are marked by N<sup>I</sup> with <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3bb52a28-0e8f-4d32-9842-6d9fb1fb69a8.png" xlink:type="simple"/></inline-formula> and form a straight line or chain structure. We associate</p><fig id="fig1"><label>Figure 1</label><caption><p> A maximal internal chain of length k</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\53c77ef3-8bd2-42b2-b81f-700b283b5f5a.png"/></fig><fig id="fig2"><label>Figure 2</label><caption><p> A maximal terminal chain of length k</p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a9fe54bf-68d3-46e9-81b6-75138be370f9.png"/></fig><fig id="fig3"><label>Figure 3</label><caption><p> The plumbing graph<img src="htmlimages\6-7402261x\6e599802-a3a0-456c-b834-0ee10b4a3e5c.png" width="31.25" height="37.5" /></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\54b9b8d1-2471-4f27-8c91-8acd742fd0f2.png"/></fig><p>to each node a weight equal to zero that is connected with using of the unnormalized Seifert invariants for Bh-spheres, which are the block elements for the construction of graph manifolds [<xref ref-type="bibr" rid="scirp.47626-ref6">6</xref>] .</p><p>Now let’s define a Laplacian matrix for the plumbing graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0b80a19e-5b3f-4599-bd32-7b9f964f53af.png" xlink:type="simple"/></inline-formula> as follows:</p><disp-formula id="scirp.47626-formula357"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\72a90a4a-f839-4b7d-95d7-4103fb3cf4de.png"/></disp-formula><p>with integer numbers <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\170d0482-61cd-4912-b8be-9f483589fcd0.png" xlink:type="simple"/></inline-formula> corresponding to each vertex<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\7b2a10c1-a088-4abf-811a-029d9207c38b.png" xlink:type="simple"/></inline-formula>. This is a tridiagonal block matrix containing all the information about the graph<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\65a54793-36ea-4260-97c1-5aaa1e1b857d.png" xlink:type="simple"/></inline-formula>. The I-th fragment of the matrix <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1032fcae-7903-46b3-873b-e0cac09d3920.png" xlink:type="simple"/></inline-formula> which corresponds to the I-th piece of the graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a84f8821-4fa8-419f-9e3d-671f6a39adb3.png" xlink:type="simple"/></inline-formula> shown in <xref ref-type="fig" rid="fig3">Figure 3</xref> is represented as</p><disp-formula id="scirp.47626-formula358"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\7f467e8d-f63e-4a9e-91dd-af3107c5c296.png"/></disp-formula><p>Note that <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\49bdf031-9b83-4b9e-ac5c-f294b3f2149f.png" xlink:type="simple"/></inline-formula> denotes an integer number 0 corresponding to the node<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d47d5c7a-773d-45b9-a8ff-5b14403133cc.png" xlink:type="simple"/></inline-formula>. Now we pay attention to the tridiagonal submatrices (blocks) of type</p><disp-formula id="scirp.47626-formula359"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\eb9f0b9b-96b2-4e48-bdc8-b6eb74a000da.png"/></disp-formula><p>and notice that using Gauss-Neumann partial diagonalization [<xref ref-type="bibr" rid="scirp.47626-ref7">7</xref>] the matrix is equivalent to the rational block matrix</p><disp-formula id="scirp.47626-formula360"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\54bb11f4-eb64-4a68-a0d2-6427991c7800.png"/></disp-formula><p>where</p><disp-formula id="scirp.47626-formula361"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\bd14948c-0247-4eb7-88ec-25a4bc365b12.png"/></disp-formula><p>and<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\07a09394-7367-4378-8e2e-270f17920807.png" xlink:type="simple"/></inline-formula>, <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a041b310-fda5-4185-b33e-6a7a14cb1ca8.png" xlink:type="simple"/></inline-formula>,<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\555c293b-8eb9-43f6-b94f-8bc24117338c.png" xlink:type="simple"/></inline-formula>. Here we are using the standard definition of continued fraction</p><disp-formula id="scirp.47626-formula362"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\fadd25c6-a4ab-4658-9335-a98033d2b381.png"/></disp-formula><p>Applying the general Gauss-Neumann partial diagonalization method for the matrix <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\206bbd5b-b96d-41eb-8854-8149c2e27998.png" xlink:type="simple"/></inline-formula> we obtain a similar result where <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\bf36af69-eaf8-4987-ad4b-4dc5ef00c7c7.png" xlink:type="simple"/></inline-formula> is a rational tridiagonal matrix of rank <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\4a80601f-59e8-488c-82b7-e7a78b541ed1.png" xlink:type="simple"/></inline-formula> (the number of nodes of the graph<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\53fb5f65-83cf-429e-a573-35e7222f6072.png" xlink:type="simple"/></inline-formula>) whose elements on the diagonal are a sum of three terms representing each maximal chain connecting to the node.</p><disp-formula id="scirp.47626-formula363"><label>(1)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5dfe4786-a7e7-401d-98b4-81277fb2c5ad.png"/></disp-formula><p>where <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5abf24fc-ff8c-4991-b152-ae78f312049b.png" xlink:type="simple"/></inline-formula> are continued fractions for a terminal chain, and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1bc972bd-bdf2-47d6-837a-3eb1c9e588a9.png" xlink:type="simple"/></inline-formula> for a internal</p><p>chain. We have used the notation <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\658f454a-fe80-4d6e-8d4e-2f18a929eb99.png" xlink:type="simple"/></inline-formula> to indicate that the order of the numbers on the continued fraction is in-</p><p>verse, i.e.<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\091aa02d-3982-477c-af00-112355104493.png" xlink:type="simple"/></inline-formula>. So, we can reduce each chain of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f99a71a6-fe6d-47a8-b49d-9b950c1c0009.png" xlink:type="simple"/></inline-formula> to a rational number<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b09fc69c-82cf-483d-ba3f-d1e6c844df37.png" xlink:type="simple"/></inline-formula>, <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1964b450-4a59-4912-88ad-cc886e6816cf.png" xlink:type="simple"/></inline-formula>or <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d02e1725-8369-49e4-8e96-acad5ee2f7ab.png" xlink:type="simple"/></inline-formula></p><p>which is represented as a continued fraction, and thus reduce the original block tridiagonal matrix <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\71f06759-1763-4c27-97f0-55c4a5152db7.png" xlink:type="simple"/></inline-formula> to a tridiagonal matrix <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ee3026f8-940b-4b01-a2cc-a4cf59593e41.png" xlink:type="simple"/></inline-formula> whose size depends just of the number of nodes of<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\8337a5b0-6b07-493b-a6c1-eb1818628847.png" xlink:type="simple"/></inline-formula>. It is important to note that it is possible to obtain the original matrix <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\abe93c95-42fe-466e-88e2-57f4fa2b79a8.png" xlink:type="simple"/></inline-formula> from the reduced matrix<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1647254e-f6c1-4e63-b2b3-c664f21e8b02.png" xlink:type="simple"/></inline-formula>.</p></sec><sec id="s3"><title>3. Rational Linking Matrix for Graph Manifold <img src="htmlimages\6-7402261x\bc2b10c6-fb3e-471e-aa91-bd20be2214ed.png" width="103.75" height="56.25" /></title><p>In this section we will construct a plumbing graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\01024f97-4c4c-4cb0-a433-fb74e0e55afc.png" xlink:type="simple"/></inline-formula> codified by the same graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a1e71d11-6218-40e6-bccc-97e94277f7c8.png" xlink:type="simple"/></inline-formula> as in section 2. Now we see the weight <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\cc7d5a55-ee8c-4147-b4db-80bf866e9fb1.png" xlink:type="simple"/></inline-formula> as the Euler number of the principal S<sup>1</sup>-(U(1)-)bundle, corresponding to i-th vertex<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\69cc9a3e-772d-4f0a-a0f4-6b4ce0e5147c.png" xlink:type="simple"/></inline-formula>. We define the bundle <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5ab9bd6a-89e1-467f-8624-69bf4fe396a1.png" xlink:type="simple"/></inline-formula> associated to each vertex <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\32479ee3-dddd-49d2-8cd0-f571ee676d98.png" xlink:type="simple"/></inline-formula> as S<sup>1</sup>-bundle over <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c2b0ea93-fea2-46ec-86be-b9b4e5535ec4.png" xlink:type="simple"/></inline-formula> with the Euler number<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\38cbf9af-91a6-497f-9c42-c6b3cb4a74ae.png" xlink:type="simple"/></inline-formula>, which can be pasted together from two trivial bundles over <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1a135edc-dbc2-4ec1-adad-44d1651f01d6.png" xlink:type="simple"/></inline-formula> as follows [<xref ref-type="bibr" rid="scirp.47626-ref8">8</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref9">9</xref>]</p><disp-formula id="scirp.47626-formula364"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1d0df027-6426-45e0-9212-e89f2cc4053e.png"/></disp-formula><p>where</p><disp-formula id="scirp.47626-formula365"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\847d35f2-bf01-481f-b67d-6f7aeb522080.png"/></disp-formula><p>Note that the above is a well known description of the lens space<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a07fc445-a855-4342-a747-ad1b1fe7fb58.png" xlink:type="simple"/></inline-formula>, so the total space of the bundle is<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\51be8f7e-90c2-49af-95ee-203d637d4cff.png" xlink:type="simple"/></inline-formula>. To perform pasting operation, which is known as plumbing between the S<sup>1</sup>-bundles, we must use the trivial bundles over annuli <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\69551345-1a3f-46d3-a1b1-997e195f2f2f.png" xlink:type="simple"/></inline-formula> where A is an annulus or twice punctured sphere<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\50a59e31-65a0-492a-a733-9a07044d5385.png" xlink:type="simple"/></inline-formula>. The manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\52037d22-4211-4268-8efa-c7905efbdb13.png" xlink:type="simple"/></inline-formula> is pasted together from the manifolds <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\55ecaf19-75f1-4653-b81e-60a244f6dfbc.png" xlink:type="simple"/></inline-formula> as follows [<xref ref-type="bibr" rid="scirp.47626-ref9">9</xref>] : whenever vertices <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1685b58e-2e7f-40bb-ac4d-ab5b2516e56f.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\432da02c-cfe5-4b11-866c-b7c70f579dfb.png" xlink:type="simple"/></inline-formula> are connected by an edge <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d0463820-47d0-4fb9-a25c-7cef76bf5521.png" xlink:type="simple"/></inline-formula> in <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\186414b6-e90d-4825-abf8-b5a02df71aad.png" xlink:type="simple"/></inline-formula> we paste a boundary component <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5beb43e9-40ee-433a-87c6-c874bfdd7672.png" xlink:type="simple"/></inline-formula> of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\48c493b2-13e9-4869-80cd-2fa69eb43c4b.png" xlink:type="simple"/></inline-formula></p><p>to a boundary component <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\7b7c7c70-c91b-48bf-bb12-b795311a9b24.png" xlink:type="simple"/></inline-formula> of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\e40c8804-cc90-4af9-b63d-b855883e5ae8.png" xlink:type="simple"/></inline-formula> by the map<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c0948697-d4ef-4628-85e9-d92e80929c37.png" xlink:type="simple"/></inline-formula>:</p><p><inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\144f0a1f-91b0-40b1-8b91-d853b80a1535.png" xlink:type="simple"/></inline-formula>so the base and fiber coordinates are exchanged under the plumbing operation. Thus the edge <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ef067f1b-39b4-469b-afd4-326578b9c249.png" xlink:type="simple"/></inline-formula> corresponds to the torus <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\e91fb2a5-2c41-47a5-8da5-d17922ef92fa.png" xlink:type="simple"/></inline-formula> along which the pieces <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\24da3e3b-684c-429c-babb-2d7db31a5321.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\6945c185-99c7-4737-b1dd-70c27cdf8b41.png" xlink:type="simple"/></inline-formula> pasted together.</p><p>For example, the plumbing of the chain shown in <xref ref-type="fig" rid="fig1">Figure 1</xref> gives us the pasting</p><disp-formula id="scirp.47626-formula366"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f493f91a-3fdd-4012-8585-5d46df452f11.png"/></disp-formula><p>where <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\dd865c03-f1ec-4933-9e82-ba473ab1a1bd.png" xlink:type="simple"/></inline-formula> is a <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\aad7ceec-b449-4fde-acab-59978638b902.png" xlink:type="simple"/></inline-formula>-bundle associating with the node<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\cda1f4e5-2821-4711-a06e-4a0d4bc18336.png" xlink:type="simple"/></inline-formula>. This chain corresponds to a Seifert fibered thick torus (homeomorphic to<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\230e1f63-5b4f-4078-a4a6-6780197db6ee.png" xlink:type="simple"/></inline-formula>) in graph manifold<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\fdd7d1ad-48bf-4ab1-8d31-ae5fba68f347.png" xlink:type="simple"/></inline-formula>. Also, the terminal maximal chain shown in <xref ref-type="fig" rid="fig2">Figure 2</xref> corresponds to a Seifert fibered solid torus (homeomorphic to<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\4859b20c-730a-4ff0-ace0-0bd5f49bd8aa.png" xlink:type="simple"/></inline-formula>) in<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f8d0c33a-183d-4e2e-a6ef-080ac2aad252.png" xlink:type="simple"/></inline-formula>. The using of graphs with nodes of valence <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5b7849b2-ddd7-4d3b-aa35-34eb675caf27.png" xlink:type="simple"/></inline-formula> (as in Section 2) corresponds to plumbing of Brieskorn homology spheres (Bh-spheres) [<xref ref-type="bibr" rid="scirp.47626-ref4">4</xref>] .</p><p>Now recall that each edge <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\668e5075-8070-41cf-a6bc-274cb366e3e7.png" xlink:type="simple"/></inline-formula> of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\841150f8-554b-4db2-89af-101d295202a4.png" xlink:type="simple"/></inline-formula> relates to the embedded torus <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\2ff7c850-974d-4432-be28-37bf999c4b46.png" xlink:type="simple"/></inline-formula> and the collection of all these tori cuts the graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ef0ede81-985e-4385-acf2-c2cd8fd6abf7.png" xlink:type="simple"/></inline-formula> into disjoint union of circle bundles over n times punctured sphere <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\27caf350-1cef-4342-9e00-e2ea2cc9c29e.png" xlink:type="simple"/></inline-formula> In general, the bundles are over compact surfaces of genus g with some boundary components, see [<xref ref-type="bibr" rid="scirp.47626-ref8">8</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref9">9</xref>] . Such a collection of tori <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\06905fe3-b4cf-438b-8ddc-ceb2b48a1bef.png" xlink:type="simple"/></inline-formula> is called a graph structure on <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\2fb7f365-4ec0-42c8-b1c4-f1996f7aed02.png" xlink:type="simple"/></inline-formula> by Waldhausen [<xref ref-type="bibr" rid="scirp.47626-ref10">10</xref>] . We want to define the Jaco-Shalen-Johannson (JSJ) graph structure <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\7fad2384-5078-4088-a3db-d7030aeb9283.png" xlink:type="simple"/></inline-formula> of the Waldhausen graph structure and to specify the corresponding JSJ-decomposition of graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\294b043a-452b-40d8-87e1-5930cb44c349.png" xlink:type="simple"/></inline-formula> on the set of Seifert fibered pieces<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\e46ce4e1-d927-4f83-b4de-d6034bbc23fc.png" xlink:type="simple"/></inline-formula>. Let us denote <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a3526175-8eac-41b4-974d-0ae5f5d366f0.png" xlink:type="simple"/></inline-formula> the set of maximal chains in the graph<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b6f2bc3b-cc15-42d2-8a06-a236c5e45277.png" xlink:type="simple"/></inline-formula>. This set can be written as a disjoint union <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\6b49f0ab-d26b-4274-b989-fd1831037456.png" xlink:type="simple"/></inline-formula> where <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5d16979f-c5d8-46e5-b8e8-67541bc2d133.png" xlink:type="simple"/></inline-formula> denote the set of interior chains and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\715b8f7c-6bcf-4caa-aa4f-ed77671a6186.png" xlink:type="simple"/></inline-formula> is the set of terminal chains. The edges of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ab1a97e8-3013-45df-949f-d7f26d23a79e.png" xlink:type="simple"/></inline-formula> contained in a chain <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\172876c5-51b4-45cd-8718-837e71eec588.png" xlink:type="simple"/></inline-formula> correspond to a set of parallel tori in<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\fa30002d-a049-4934-91c5-470300252c7c.png" xlink:type="simple"/></inline-formula>. Choose one torus <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\79a43cfa-052e-426f-acb8-6d9168df06a9.png" xlink:type="simple"/></inline-formula> among them and define</p><disp-formula id="scirp.47626-formula367"><label>(2)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0c9f25c6-ff2d-4aff-b660-ad559de0a674.png"/></disp-formula><p>This set of tori performs the well known JSJ-decomposition of the graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\12f1b3a4-2e85-4cb0-a3bb-b55218b08446.png" xlink:type="simple"/></inline-formula> [<xref ref-type="bibr" rid="scirp.47626-ref11">11</xref>] .</p><p>By construction, each piece <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\83be5f08-1cb4-46e8-a4d4-335bd04be696.png" xlink:type="simple"/></inline-formula> (denoted as <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\77f95669-1c57-4fec-9736-f14d27dfa63c.png" xlink:type="simple"/></inline-formula> for brevity) of JSJ-decomposition that corresponds to the node <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\7e3bb1e8-b6b3-4bbe-b16e-a9bb450b8842.png" xlink:type="simple"/></inline-formula> contains a unique piece <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\7ac7dec4-777f-41e4-88ba-ab256707d92c.png" xlink:type="simple"/></inline-formula> (which we shall denote as<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\839950e5-36a1-4b94-8118-39c4dd2f42f7.png" xlink:type="simple"/></inline-formula>) of Waldhausen de- composition associated with the same node<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\39d1c96b-a4c2-47a9-a086-0c6e5ed2ed9c.png" xlink:type="simple"/></inline-formula>. One can extend in a unique way up to isotopy the natural Seifert structure without exceptional fibers on <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\e752634e-9ceb-463c-87cd-e53c2186351b.png" xlink:type="simple"/></inline-formula> to a Seifert fibration on <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ad488cca-7a76-49c9-b8c8-1ddca40919d2.png" xlink:type="simple"/></inline-formula> with exceptional fibers. Thus in these terms the JSJ-decomposition of the manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\45f5c41b-8218-4a94-bd1a-073b07c1b691.png" xlink:type="simple"/></inline-formula> is defined completely by <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3923846f-b92e-4753-87e9-df058ab2e807.png" xlink:type="simple"/></inline-formula> where R is the number of nodes in <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\bc255208-450c-4c30-b957-d1f91aee28b8.png" xlink:type="simple"/></inline-formula> and the bar over M means the closure of the piece<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\873d8d7d-268a-422f-8f87-51dbb2a57b02.png" xlink:type="simple"/></inline-formula>.</p><p>Note that there exists an uncertainty in the choice of the torus <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3474de37-1673-4ae0-b0c4-8e716588f2bb.png" xlink:type="simple"/></inline-formula> for each internal chain which appear in the JSJ-structure (2). We can remove this uncertainty in following way. Let us perform the maximal extension of the natural Seifert fibration from each <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c5b9097e-c939-4e31-a1de-3b440858db6f.png" xlink:type="simple"/></inline-formula> and denote the obtained Seifert fibered piece of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b36256af-9272-4839-b86b-2d0aea95ad81.png" xlink:type="simple"/></inline-formula> by<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0202d8bd-5191-4656-bc8b-57df8bf2ffb5.png" xlink:type="simple"/></inline-formula>. It is clear that <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3eeb9088-29ef-40e8-8be2-967eb06a863e.png" xlink:type="simple"/></inline-formula> if and only if there exists a chain <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\aec3ccaf-da4b-464a-ad98-8f3433b89cc5.png" xlink:type="simple"/></inline-formula> joining the nodes <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f934b89b-f9e8-4dac-ac07-551045ff8e47.png" xlink:type="simple"/></inline-formula> and<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3b72d82e-725d-4320-b26b-ab868b3eb116.png" xlink:type="simple"/></inline-formula>. If we start with plumbing of R Bh-spheres<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\9c206596-0182-4171-b58f-503731445076.png" xlink:type="simple"/></inline-formula>, the resulting graph three-manifold will be integer homology sphere [<xref ref-type="bibr" rid="scirp.47626-ref4">4</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref9">9</xref>] (<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\feb1abd8-473a-49e8-8bf8-473fd74a47b9.png" xlink:type="simple"/></inline-formula>-homology sphere), which in general case does not have the global Seifert fibration. But we can construct the JSJ-covering<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\45eb489e-e162-477e-b6a7-fbff0bbbae8e.png" xlink:type="simple"/></inline-formula>, such as each <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\25a2f66d-4e4d-47ee-b0e4-d6a365338eee.png" xlink:type="simple"/></inline-formula> is a Seifert fibered space and it is maximal in the sense described above.</p><p>Suppose that we perform the plumbing operation according to the plumbing diagram<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\6e8c8b5e-cd9c-4b52-b4de-81ce01765f5a.png" xlink:type="simple"/></inline-formula>, shown on the <xref ref-type="fig" rid="fig4">Figure 4</xref>. Thus our plumbing diagrams will always have the pairwise coprime weights around each node and correspond to <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d4c72fee-f9a0-4850-988b-6072d23ad489.png" xlink:type="simple"/></inline-formula>-homology spheres [<xref ref-type="bibr" rid="scirp.47626-ref5">5</xref>] .</p><p>We construct the plumbing graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\fd968ca7-18a8-44d2-9a85-8ef36462d1e7.png" xlink:type="simple"/></inline-formula> for a <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\78703675-120f-48c2-9285-b6aaa3ea580d.png" xlink:type="simple"/></inline-formula>-homology sphere, following [<xref ref-type="bibr" rid="scirp.47626-ref5">5</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref8">8</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref9">9</xref>] (as a result we shall obtain a graph of type shown in <xref ref-type="fig" rid="fig3">Figure 3</xref>). First of all we calculate the characteristics of maximal chains. For terminal chains the integer Euler numbers <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\6c53b9f4-5dd3-45de-bceb-ad2d3aa70b69.png" xlink:type="simple"/></inline-formula> are defined by the continued fraction:</p><fig id="fig4"><label>Figure 4</label><caption><p> A plumbing (splicing) diagram Δ<sub>p</sub></p></caption><graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a47b3816-cb28-4bdb-b791-b096fbaccdc7.png"/></fig><disp-formula id="scirp.47626-formula368"><label>(3)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\4c08975d-4031-4121-9da6-a60bd2d0637b.png"/></disp-formula><p>where<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f635b2ae-33cd-4caa-97b1-45f97b26ce5c.png" xlink:type="simple"/></inline-formula>, <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\61052808-3222-4155-8ed8-014bf126d79d.png" xlink:type="simple"/></inline-formula>are the Seifert invariants, numerated in the following way</p><disp-formula id="scirp.47626-formula369"><label>(4)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3d56cd8c-273e-4335-adab-bab22c980ec2.png"/></disp-formula><p>For internal chains the integer Euler numbers <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\2fefdfea-25d8-4ae1-a48f-0d4da6cd6871.png" xlink:type="simple"/></inline-formula> are defined by</p><disp-formula id="scirp.47626-formula370"><label>(5)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b2b05d5f-2315-47c7-9807-8a5ed96fd75b.png"/></disp-formula><p>where the Seifert (orbital) invariants<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b36b9562-a6b5-40ac-9533-8b5a42455ec3.png" xlink:type="simple"/></inline-formula>, <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\8ab8c8cc-a5c6-4742-9b74-49fc7e3dccc3.png" xlink:type="simple"/></inline-formula>characterize the thick tori<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\364f1315-9d18-4f1a-b76f-e0896876330c.png" xlink:type="simple"/></inline-formula>, which are created by the plumbing operations performed between the nodes <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a3ba0655-27cc-4e3e-ae33-0820658d4804.png" xlink:type="simple"/></inline-formula> and<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c1cb118d-039d-4743-afd1-b5a4eeee7a4f.png" xlink:type="simple"/></inline-formula>, see [<xref ref-type="bibr" rid="scirp.47626-ref6">6</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref9">9</xref>] . These invariants identify also the extra lens spaces <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3587d98c-fa8f-4da0-bd8d-68022ed263bc.png" xlink:type="simple"/></inline-formula> which arise in four-dimensional plumbed V-cobordism (corresponding to the graph<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\bbffe1d3-9008-4325-946f-bac24fdbdfb7.png" xlink:type="simple"/></inline-formula>) with</p><disp-formula id="scirp.47626-formula371"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\80436279-8b5f-4978-982f-278f50ff32e3.png"/></disp-formula><p>for the ordering fixed by the plumbing diagram in <xref ref-type="fig" rid="fig4">Figure 4</xref>. From this representation of the plumbing graph it is clear that for <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\4657e160-4fd9-4fac-84ad-22d722e17735.png" xlink:type="simple"/></inline-formula> the set <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\31fd451f-c2bb-4a32-8fc4-72bf6152cff0.png" xlink:type="simple"/></inline-formula> of JSJ-covering <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\caab8297-0346-44e0-9bcf-46a55aafa7c5.png" xlink:type="simple"/></inline-formula> has the form</p><disp-formula id="scirp.47626-formula372"><label>(6)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5db92ce3-7db5-4fe0-bae6-e343cfc684af.png"/></disp-formula><p>where <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\2e21b0d8-6087-4a6e-aa8e-4929fa8d313c.png" xlink:type="simple"/></inline-formula> is a Seifert fibered solid torus with Seifert invariants <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\4550eb99-c335-4ca7-992e-e5044216bd3c.png" xlink:type="simple"/></inline-formula> and</p><disp-formula id="scirp.47626-formula373"><label>(7)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\550a0a0d-e995-43d4-96cc-77736c66cf94.png"/></disp-formula><p>For the cases <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b91a37f1-cc60-4191-9071-b9d5388e7839.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d33a8025-59fd-47f6-b746-1f69a894227c.png" xlink:type="simple"/></inline-formula> the formulas are different from (6):</p><disp-formula id="scirp.47626-formula374"><label>(8)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\9e9e6077-1c63-48e2-9477-8d3451fbf95c.png"/></disp-formula><disp-formula id="scirp.47626-formula375"><label>(9)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c95c7109-ecf1-4866-88ae-d09b6cf26d4f.png"/></disp-formula><p>Moreover <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\db2f8a15-97d6-4d55-96a6-cc82c7d8e248.png" xlink:type="simple"/></inline-formula> Here the symbol <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\51233f9c-91f4-49ea-a674-e65bc3eb6519.png" xlink:type="simple"/></inline-formula> indicates that  <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\65c1fbd6-257b-4f46-9d13-edeae6720e94.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\4eaaeb6b-1cbc-497f-949b-32c746b96286.png" xlink:type="simple"/></inline-formula> are homeomorphic, but their Seifert structures are characterized by different integer Euler numbers defined by (5) and (7) respectively. Thereby the thick torus between the nodes N<sup>I</sup> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0a566fe9-9b23-4669-92c9-f374dc6cf8d7.png" xlink:type="simple"/></inline-formula> has two Seifert fibrations: the first is the extension of the natural Seifert fibration defined on the piece <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\57076b8e-1551-41f8-a5b8-054aebcc20bf.png" xlink:type="simple"/></inline-formula> and the second one is obtained as extension from the piece<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\9d0ccf4a-0ec5-449b-a52c-bafc93978d49.png" xlink:type="simple"/></inline-formula>. These Seifert fibrations are connected by the matrix [<xref ref-type="bibr" rid="scirp.47626-ref6">6</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref9">9</xref>] :</p><disp-formula id="scirp.47626-formula376"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\90206088-4b9b-4b9c-ada4-35bec1ee545c.png"/></disp-formula><p>in the following sense. Recall that edges of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\710d7e22-3f5a-4e6b-abf9-853945be077f.png" xlink:type="simple"/></inline-formula> contained in a chain <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d3e1a510-5a19-4d81-a7ac-edcea4d585e1.png" xlink:type="simple"/></inline-formula> (between the nodes N<sup>I</sup> and<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0efa262a-0aa3-42f4-92c1-80137cdc4d89.png" xlink:type="simple"/></inline-formula>) correspond to the set parallel tori in<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\34a85b58-5ea7-4fe1-b028-94046c86214c.png" xlink:type="simple"/></inline-formula>. On any of these tori there exist two bases formed by the section lines and the fibers pertain to the Seifert fibrations extended from <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\cd3340de-84bc-4e72-b2db-4f29b9d0b1f5.png" xlink:type="simple"/></inline-formula> and<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\565a464a-0cde-4329-9aee-efb1f5ef8f32.png" xlink:type="simple"/></inline-formula>, which we denote as the pair of columns</p><disp-formula id="scirp.47626-formula377"><label>(10)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\bf85875a-1695-483f-9c80-f8852480536f.png"/></disp-formula><p>Subindices 2 and 3 manifest that <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\a6c01e06-da8f-42be-9640-22b40c912d9e.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\918e2756-7f4b-4120-aa39-87271403ad76.png" xlink:type="simple"/></inline-formula> are plumbed together along the singular fibers with Seifert invariants <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0dceb89e-de8e-4bc9-9c61-d6341a19bfcf.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ee0dccb3-ee35-41c1-acfb-a353fe7b91dc.png" xlink:type="simple"/></inline-formula> (see <xref ref-type="fig" rid="fig4">Figure 4</xref>). Then the transformation between these section-fiber bases is described by</p><disp-formula id="scirp.47626-formula378"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\dfd9ce01-2698-4486-926e-304a9c54b447.png"/></disp-formula><p>where <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3ae08adb-065f-4712-8c9f-bfa339b1d128.png" xlink:type="simple"/></inline-formula> is defined from<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0f655b3b-c131-421a-a996-69ce3b22fd82.png" xlink:type="simple"/></inline-formula>.</p><p>Now we introduce the one-form bases <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d277275e-816c-4740-8a79-4fc6b40caf37.png" xlink:type="simple"/></inline-formula> and<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c7fd6168-9442-444d-89a4-2b8e3333799e.png" xlink:type="simple"/></inline-formula>, duals to the bases (10) in the following sense:</p><disp-formula id="scirp.47626-formula379"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\036b755d-0b68-4d90-9879-fd7d877db290.png"/></disp-formula><p>where the integrals are calculated over any such section line or fiber as, for example, in [<xref ref-type="bibr" rid="scirp.47626-ref12">12</xref>] . Thus we obtain the corresponding transformations between the the dual one-forms:</p><disp-formula id="scirp.47626-formula380"><label>(11)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\9bfba54d-45c8-4f9b-9aa5-85c2c52d9fa9.png"/></disp-formula><p>We suppose that the forms <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\bd13b8f8-046c-4647-9eae-46ed82b95a67.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0a569062-0066-4397-a10c-c0a569e59426.png" xlink:type="simple"/></inline-formula> are dual with respect to the bilinear pairing defined as</p><disp-formula id="scirp.47626-formula381"><label>(12)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\32fe1154-92e8-4949-812c-96b8e474076e.png"/></disp-formula><p>Also we shall used the integrals</p><disp-formula id="scirp.47626-formula382"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1d7617d8-ce5a-46bf-b6ec-0548bb17e2b7.png"/></disp-formula><p>which define the linking (intersection) numbers of the fiber structures <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\fe65898b-e0b9-4813-bf5a-582e86821b05.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\aff5e9ab-d76d-4b32-b92b-d3f46b0929ce.png" xlink:type="simple"/></inline-formula> defined on thick torus<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\fffd7400-4666-44f0-8c36-0a2acb86a383.png" xlink:type="simple"/></inline-formula>. We can obtain the rational linking matrix for <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\fcf78502-1a40-468b-8eaa-8233134ae161.png" xlink:type="simple"/></inline-formula> by means of multiplication of the Equations (11) by <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\1a6c67c0-0182-49f2-bd6d-de17da2f95c2.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5a013428-cae3-4f65-84ec-27a2c4b6af1d.png" xlink:type="simple"/></inline-formula> and integration over<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\46a45dcb-b461-41f7-94f2-c9d692af2613.png" xlink:type="simple"/></inline-formula>. Applying the duality conditions (12) we obtain:</p><disp-formula id="scirp.47626-formula383"><label>(13)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\9cdfed21-6643-4e16-b776-d68d490d5177.png"/></disp-formula><p>The rational numbers <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f048ad04-b79e-4605-9853-adb6edbc8c48.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c28d8d8b-85fc-4e58-b1a3-ebc46e6fc73b.png" xlink:type="simple"/></inline-formula> are also known as Chern classes of the line V-bundles associated with the Seifert fibrations with the U(1)-invariant connection forms <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\0fd9ac81-a31f-4bc6-a771-9e651e8a7f7b.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\2c251f1f-4232-4017-a79a-c8c4e09ae607.png" xlink:type="simple"/></inline-formula> on the lens spaces <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\bf253b97-0804-41d6-a852-3c8346fbce1a.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\14c3568e-fe00-4143-9286-07585b71c9db.png" xlink:type="simple"/></inline-formula> respectively [<xref ref-type="bibr" rid="scirp.47626-ref12">12</xref>] .</p><p>Now we are ready to calculate the rational linking matrix for the graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\8cddf5a9-5fdf-48e8-909e-fcbc50f3ed8c.png" xlink:type="simple"/></inline-formula> (see <xref ref-type="fig" rid="fig3">Figure 3</xref>):</p><disp-formula id="scirp.47626-formula384"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\623cf877-e9cc-4568-8982-16666fe5c759.png"/></disp-formula><p>We integrate here over the three dimensional graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\71b69407-2c4b-41ee-8a93-72235e68b855.png" xlink:type="simple"/></inline-formula> to obtain a positive definite linking matrix. This manifold has the opposite orientation with respect to the graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\13267827-a898-41c2-89fa-875525d831f6.png" xlink:type="simple"/></inline-formula> obtained directly by plumbing of Bh-spheres, which are defined as links of singularities. This construction of the graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\226fb3ee-b7a3-41c2-93e4-81dae9cd9d92.png" xlink:type="simple"/></inline-formula> gives the possibility to represent it also as a link of singularity that guarantees its rational linking matrix to be negative definite (for details see [<xref ref-type="bibr" rid="scirp.47626-ref5">5</xref>] ).</p><p>From the tree structure of the graph<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b631aebd-db1b-4749-aee6-92a2c0ea971b.png" xlink:type="simple"/></inline-formula>, and from the first equation in (13) we immediately obtain, that for <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\f5cbdb4f-ff85-4e03-90c4-2edb7815f8fb.png" xlink:type="simple"/></inline-formula> the nonzero elements are only</p><disp-formula id="scirp.47626-formula385"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3d3c4482-9957-4fe3-9c4e-66bd515df7bb.png"/></disp-formula><p>for <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5d300402-35c4-44d8-af7a-758567caddd1.png" xlink:type="simple"/></inline-formula></p><p>If<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\844d00f9-c22a-4963-902b-f31a286326cf.png" xlink:type="simple"/></inline-formula>, we have</p><disp-formula id="scirp.47626-formula386"><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\df4590d4-e620-4384-9911-baf4cf35e15f.png"/></disp-formula><p>Here we use the decomposition (6) of the piece<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ba169f1a-dae0-4055-b723-8ce9663f06f0.png" xlink:type="simple"/></inline-formula>, and that the integral over trivial Seifert fibration <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\2c8ca377-1e3b-4b7d-aaa8-4463cd450428.png" xlink:type="simple"/></inline-formula> is zero. Then according to the two last equations in (13) we obtain the matrix element</p><disp-formula id="scirp.47626-formula387"><label>(14)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5625d67d-8f00-443d-8f6e-15ebcf46057f.png"/></disp-formula><p>also known as the Chern class of line V-bundle associated with the Seifert fibration of<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\b30ad791-8e6d-4ce0-a1d2-8754ec91d20e.png" xlink:type="simple"/></inline-formula>.</p><p>For <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c028cc76-17d1-4782-8ac4-9c45e7bd48e6.png" xlink:type="simple"/></inline-formula> and <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\76bd087e-42a8-4cad-b055-e71bbed27299.png" xlink:type="simple"/></inline-formula> the matrix elements are</p><disp-formula id="scirp.47626-formula388"><label>(15)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\845de791-4333-4634-8f66-4612d4483d31.png"/></disp-formula><disp-formula id="scirp.47626-formula389"><label>(16)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\3dfc184d-e563-467c-8808-a2fe3efadf07.png"/></disp-formula><disp-formula id="scirp.47626-formula390"><label>(17)</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\79a00d32-cc92-4e53-a972-d30c3e8ca504.png"/></disp-formula><p>Here we have used the decompositions (8) and (9) as well as the notations (4).</p></sec><sec id="s4"><title>4. Conclusions</title><p>Comparing the reduced matrix <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\df48a03a-1f48-4331-a720-f5ea45cd244a.png" xlink:type="simple"/></inline-formula> (1) with the results (14) and (17) for the rational linking matrix K of the graph manifold <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\73f92c60-4327-42dd-b346-c8f3cf1759cb.png" xlink:type="simple"/></inline-formula> we observe that decomposing the rational invariants into continued fractions according to (3), (5) and (7), we can create the graph <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\aec74405-cd8e-4dbc-bc72-4f43b94bb4c2.png" xlink:type="simple"/></inline-formula> (related to diagram<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c6238a96-d05d-429f-aca9-a44ba19badc5.png" xlink:type="simple"/></inline-formula>) and obtain the rational linking matrix K of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\85dff707-5397-42bb-80e6-63530bee4700.png" xlink:type="simple"/></inline-formula> just by Gauss-Neumann diagonalization on the Laplacian matrix of<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\5448a3a6-0386-455f-bb20-8fd3c890391f.png" xlink:type="simple"/></inline-formula>. This is the main result of this article.</p><p>We want to conclude with an example of an application of our results for the topological field theory. In [<xref ref-type="bibr" rid="scirp.47626-ref3">3</xref>] [<xref ref-type="bibr" rid="scirp.47626-ref13">13</xref>] we built a set of graph manifolds whose Seifert invariants are constructed on the base of the first 9 prime numbers<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\d10624da-653c-4485-a2a7-508f443b3da0.png" xlink:type="simple"/></inline-formula>. The rational linking matrix of these graph manifolds are positive definite and have diagonal elements (and eigenvalues) simulating the low-energy coupling constants hierarchy of the fundamental interactions of real universe. An example of such matrix is [<xref ref-type="bibr" rid="scirp.47626-ref13">13</xref>]</p><disp-formula id="scirp.47626-formula391"><label>,</label><inline-graphic xmlns:xlink="http://www.w3.org/1999/xlink" xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\49000f88-1a89-4c6f-af2c-934aeb804bad.png"/></disp-formula><p>whose elements are all rational and the diagonal ones are described in (14) and (17) by a sum of three continued fractions. The matrix <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\09e69402-e90d-4a0a-baad-e7a613de7b94.png" xlink:type="simple"/></inline-formula> inverse to <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\ca6e9a51-db91-4da8-a648-001c0d88d4aa.png" xlink:type="simple"/></inline-formula> is integer one [<xref ref-type="bibr" rid="scirp.47626-ref6">6</xref>] , the inversion of the rational linking matrices can be done with the help of any program such as MathematicaTM to verify this property, any error on calculation of <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\c5324993-ebf5-47ef-b0f4-d8e71c2d28a1.png" xlink:type="simple"/></inline-formula> leads to non-integer elements in resulting matrices<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\060f4e20-27cb-4d8c-bb0e-debc04286bd2.png" xlink:type="simple"/></inline-formula>. It is also worth mentioning that in the present example the Laplacian block matrix<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\96ff56a6-39e9-45b4-9552-2e734fad0a29.png" xlink:type="simple"/></inline-formula>, corresponding to the matrix<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\302bfa2a-cad7-47ba-80f2-ae91f54255ab.png" xlink:type="simple"/></inline-formula>, has  <inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\6dd184fc-f6f5-4c2e-a497-f175f18d71ec.png" xlink:type="simple"/></inline-formula> [<xref ref-type="bibr" rid="scirp.47626-ref13">13</xref>] , while<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\e958132b-37b4-4481-8ae4-7066c2ea578e.png" xlink:type="simple"/></inline-formula>.</p><p>In the 7-dimensional Kaluza-Klein approach to the topological field theory (BF-model), the rational linking matrices of the 3-dimensional graph manifold may be really interpreted as coupling constants matrices [<xref ref-type="bibr" rid="scirp.47626-ref1">1</xref>] . So, the Gauss-Neumann method gives us a simple algorithm to calculate the linking matrices of graph manifolds and thus the coupling constants matrices (despite the probably huge rank of the original block matrix<inline-formula><inline-graphic xlink:href="http://file.scirp.org/Html/htmlimages\6-7402261x\52bf8580-a022-40b6-b506-aa727501d240.png" xlink:type="simple"/></inline-formula>).</p></sec></body><back><ref-list><title>References</title><ref id="scirp.47626-ref1"><label>1</label><mixed-citation publication-type="journal" xlink:type="simple"><name name-style="western"><surname>VERLINDE</surname><given-names> E. </given-names></name>,<etal>et al</etal>. (<year>1995</year>)<article-title>GLOBAL ASPECTS OF ELECTRIC-MAGNETIC DUALITY</article-title><source>. NUCLEAR PHYSICS B</source><volume> 455</volume>,<fpage> 211</fpage>-<lpage>225</lpage>.<pub-id pub-id-type="doi">HTTP://DX.DOI.ORG/10.1016/0550-3213(95)00431-Q</pub-id></mixed-citation></ref><ref id="scirp.47626-ref2"><label>2</label><mixed-citation publication-type="journal" xlink:type="simple"><name name-style="western"><surname>EFREMOV</surname><given-names> V.N.</given-names></name>,<name name-style="western"><surname> MITSKIEVICH</surname><given-names> N.V.</given-names></name>,<name name-style="western"><surname> HERN</surname><given-names>EZ MAGDALENO</given-names></name>,<name name-style="western"><surname> A.M. </surname><given-names> SERRANO BAUTISTA</given-names></name>,<name name-style="western"><surname> R. </surname><given-names>  </given-names></name>,<etal>et al</etal>. (<year>2005</year>)<article-title>TOPOLOGICAL GRAVITY ON PLUMBED V-COBORDISMS</article-title><source>. CLASSICAL AND QUANTUM GRAVITY</source><volume> 22</volume>,<fpage> 3725</fpage>-<lpage>3744</lpage>.<pub-id pub-id-type="doi">HTTP://DX.DOI.ORG/10.1088/0264-9381/22/17/022</pub-id></mixed-citation></ref><ref id="scirp.47626-ref3"><label>3</label><mixed-citation publication-type="journal" xlink:type="simple"><name name-style="western"><surname>EFREMOV</surname><given-names> V.N.</given-names></name>,<name name-style="western"><surname> HERN</surname><given-names>EZ MAGDALENO</given-names></name>,<name name-style="western"><surname> A.M. </surname><given-names> MORENO</given-names></name>,<name name-style="western"><surname> C. </surname><given-names>  </given-names></name>,<etal>et al</etal>. (<year>2010</year>)<article-title>TOPOLOGICAL ORIGIN OF THE COUPLING CONSTANTS HIERARCHY IN KALUZA-KLEIN APPROACH</article-title><source>. INTERNATIONAL JOURNAL OF MODERN PHYSICS A</source><volume> 25</volume>,<fpage> 2699</fpage>-<lpage>2733</lpage>.<pub-id pub-id-type="doi">HTTP://DX.DOI.ORG/10.1142/S0217751X10048482</pub-id></mixed-citation></ref><ref id="scirp.47626-ref4"><label>4</label><mixed-citation publication-type="other" xlink:type="simple">SAVELIEV, N. (2002) INVARIANTS FOR HOMOLOGY3-SPHERES. SPRINGER, BERLIN, 223. HTTP://DX.DOI.ORG/10.1007/978-3-662-04705-7</mixed-citation></ref><ref id="scirp.47626-ref5"><label>5</label><mixed-citation publication-type="other" xlink:type="simple">EISENBUD, D. AND NEUMANN, W. (1985) THREE-DIMENSIONAL LINK THEORY AND INVARIANTS OF PLANE CURVE SINGULARITIES. PRINCETON UNIVERSITY PRESS, PRINCETON, 172.</mixed-citation></ref><ref id="scirp.47626-ref6"><label>6</label><mixed-citation publication-type="journal" xlink:type="simple"><name name-style="western"><surname>SAVELIEV</surname><given-names> N. </given-names></name>,<etal>et al</etal>. (<year>2002</year>)<article-title>FUKUMOTO-FURUTA INVARIANTS OF PLUMBED HOMOLOGY 3-SPHERES</article-title><source>. PACIFIC JOURNAL OF MATHEMATICS</source><volume> 205</volume>,<fpage> 465</fpage>-<lpage>490</lpage>.<pub-id pub-id-type="doi">HTTP://DX.DOI.ORG/10.2140/PJM.2002.205.465</pub-id></mixed-citation></ref><ref id="scirp.47626-ref7"><label>7</label><mixed-citation publication-type="journal" xlink:type="simple"><name name-style="western"><surname>NEUMANN</surname><given-names> W. </given-names></name>,<etal>et al</etal>. (<year>1997</year>)<article-title>COMMENSURABILITY AND VIRTUAL FIBRATION FOR GRAPH MANIFOLDS</article-title><source>. TOPOLOGY</source><volume> 36</volume>,<fpage> 355</fpage>-<lpage>378</lpage>.<pub-id pub-id-type="doi">HTTP://DX.DOI.ORG/10.1016/0040-9383(96)00014-6</pub-id></mixed-citation></ref><ref id="scirp.47626-ref8"><label>8</label><mixed-citation publication-type="other" xlink:type="simple">HIRZEBRUH, F. (1971) DIFFERENTIABLE MANIFOLDS AND QUADRATIC FORMS. MARCEL DEKKER, NEW YORK, 56-58.</mixed-citation></ref><ref id="scirp.47626-ref9"><label>9</label><mixed-citation publication-type="other" xlink:type="simple">NEUMANN, W. (1981) A CALCULUS FOR PLUMBING APPLIED TO THE TOPOLOGY OF COMPLEX SURFACE SINGULARITIES AND DEGENERATING COMPLEX CURVES. TRANSACTIONS OF THE AMERICAN MATHEMATICAL SOCIETY, 268, 299. HTTP://DX.DOI.ORG/10.1090/S0002-9947-1981-0632532-8</mixed-citation></ref><ref id="scirp.47626-ref10"><label>10</label><mixed-citation publication-type="journal" xlink:type="simple"><name name-style="western"><surname>WALDHAUSEN</surname><given-names> F. </given-names></name>,<etal>et al</etal>. (<year>1967</year>)<article-title>EINE KLASSE VON 3-DIMENSIONALEN MANNIGFALTIGKEITEN. I</article-title><source>. INVENTIONES MATHEMATICAE</source><volume> 3</volume>,<fpage> 308</fpage>-<lpage>333</lpage>.<pub-id pub-id-type="doi">HTTP://DX.DOI.ORG/10.1007/BF01402956</pub-id></mixed-citation></ref><ref id="scirp.47626-ref11"><label>11</label><mixed-citation publication-type="other" xlink:type="simple">POPESCU-PAMPU, P. (2007) THE GEOMETRY OF CONTINUED FRACTIONS AND THE TOPOLOGY OF SURFACE SINGULARITIES. IN: BRASSELET, J.-P. AND SUWA, T., EDS., SINGULARITIES IN GEOMETRY AND TOPOLOGY 2004, ADVANCED STUDIES IN PURE MATHEMATICS, VOL. 46, 119-195.</mixed-citation></ref><ref id="scirp.47626-ref12"><label>12</label><mixed-citation publication-type="journal" xlink:type="simple"><name name-style="western"><surname>BEASLEY</surname><given-names> C. </given-names></name>,<name name-style="western"><surname> WITTEN</surname><given-names> E. </given-names></name>,<etal>et al</etal>. (2005)<article-title>BEASLEY, C. AND WITTEN, E.  NON-ABELIAN LOCALIZATION FOR CHERN-SIMONS THEORY</article-title><source>. JOURNAL OF DIFFERENTIAL GEOMETRY</source><volume> 70</volume>,<fpage> 183</fpage>-<lpage>323</lpage>.<pub-id pub-id-type="doi"></pub-id></mixed-citation></ref><ref id="scirp.47626-ref13"><label>13</label><mixed-citation publication-type="other" xlink:type="simple">EFREMOV, V., HERNANDEZ, A. AND BECERRA, F. (2014) THE UNIVERSE AS A SET OF TOPOLOGICAL FLUIDS WITH HIERARCHY AND FINE TUNING OF COUPLING CONSTANTS IN TERMS OF GRAPH MANIFOLDS. ARXIV:1309.0690V2</mixed-citation></ref></ref-list></back></article>