# Existing 214: uniform existence and actual completeness interfaces Author: `/root/d214_uniform_existence`, 2026-11-11. Author argument only; fresh non-author acceptance is required. No source, old packet, coefficient, witness table or public repository is changed. This is an additive clarification of the frozen existing 214 family, with its original hypotheses retained exactly. Source bindings and actual consumer passages are in SOURCE_USES.json and CONTRACT.json. [P01: exact theorem] Let the 214 family be exactly the family with 188 primitive integral quadratic residuals in `research/quadratic-ten-contact-integration/existence/artifacts/candidate214-brox/`, at private HEAD bfeb8400982c9a169e4fdd39fb68ec053621e8ed. Its parameter is the fixed coefficient A_N=149^(2N+1), not an existential coordinate. Subject to the named full-family height and retained elliptic/rank interfaces below, the conclusion is exists N0 in Z_{>=1}, for every integer N>=N0, for every x in Q, x in Z iff there exists y in Q^213 with P_N(x,y)=0. This is the same polynomial family. N0 is not computed. N>=53 is a coefficient-height condition only; it need not be included in the arithmetic N0, and replacing N0 by max(N0,53) is harmless. No N=53 defining theorem or numerical defining member follows. [P02: scope of source carry] The entire latest PR26 manuscript at public checkout HEAD 1a49013b5683f58d2f88bb018a07d13770354157 and the old 214 frozen `existence/sources/quartic_v15.tex` have the identical **raw byte** SHA256 44ba3a8ec8521c29b81b26cfb9981cd6d5f3c76d54e7782438614842edb52c74 and 1924438 bytes. This is stronger than normalized-line identity. Thus no analytic, restricted-converse, aggregate, scalar or prime-selection manuscript repair is newly imported from PR26 here. The later audit records can justify exact unchanged applications at their stated scope only after source/consumer comparison; PR status is not an arithmetic premise. This author read the actual source passages identified in SOURCE_USES, including the sum-pole/graph argument, the prime theorem/obstruction-tail/fixed-field and diagonal statements, the CRT argument, scalar-S/four-contact preparation and zero-preparation proof. This author did not reconstruct every internal RC/analytic sublemma or final published source anew. Those unchanged internal foundations and their eligible source review remain a separately bound dependency; this packet closes the 214 application joins, not their entire recursive review. [P03: exact retained inputs] Use K=Q(sqrt2), d=3-sqrt2, E:y^2=x^3-d^2 x, and the same P=2(1+sqrt2,2sqrt2). The four nonzero integer labels are (n,a,t,b), c=|a| and B=max(|n|,|a|,|t|,|b|), so B>=c. The physical scale is T=(1/4)Tr(sum_i x([n_i]P))>0. The physical arguments are t_0=lambda and t_1=T lambda. The contact set is Theta_2=Theta union(2-Theta), Theta={0,-1/2,-1,1/2399,-32/407}; Sigma is the primes<=76176 together with 2029147. Each comparison radical M_j is the squarefree product of primes p outside Sigma at a positive odd contact of t_j. Do not substitute the sibling center-3/11 contact system or its input 51(2X+1)^2+2504. Here beta=51x^2+2504. [P04: restored objects and branches] The old existence proof's exact Laurent clearing and raw224 inverse maps reconstruct the physical reference/scalar/root system before any ratio is consumed. Its recovered f,S,g and Q2 coefficient at149 imply w=g^4/f^2; deleted fourth powers are their actual squares. The shared alternate ratio has numerator and denominator of valuation -v149(f)/2 and is nonzero. The unchanged scalar argument excludes reference torsion and H indices -1,0,1 before logarithms. No rational component or tested numerator is divided by. The new inverse rows enforce q_j!=0, q_j=1/R_j-1. They remove an irrelevant old q_j=0 witness stratum, so this is not an all-old-witness bijection. The stratum delta_j=t_j-1=0 has q_j=-1 and M_j=1, and needs no division by delta_j. A contact t_j in Theta_2 is excluded by F_2(t_j)(1+q_j)=delta_j^10 and F_2(1)!=0. A putative q_j=0 has radical one, but is in any event excluded in 214. None of these statements confuses delta_j with the integer Lucas residual Delta below. [P05: changed local factors] At a consumed p outside Sigma and either place above p, normalize v(p)=1. Exactly one contact factor has positive odd valuation nu, delta_j is a unit and v(q_j)=-nu. The five actual 214 factors are 1+q^5 C2^2+q^2 D_H^2+q^4 B1^4; 1+q^5 G^2+q^2 A3^2+q^4 B0^4; 1+q^5 I^4+q^3 A2^2+q^2 A4^2+q^4 B3^4; 1+q^5 C3^2+q^2 C1^4+q^4 B4^4; 1+q^5 C4^2+q^2 A1^2+q^4 B2^4. Here D_H=f(H), A_i=f(U_i), B_i=sigma f(V_i), C_i=L_i-L_i', I=L1-beta and G=L3^2-L2(L2-2)L3L4+(L2-2)^2L4^2-1, with f=1/x and f(O)=0. Every nonzero reciprocal abscissa has even valuation, including f(H-P) restored before this step. In the generic factors the finite nonconstant valuations lie in odd integers, 2+4Z and 4+8Z. In the third factor the first three have residues -nu, nu, 2 modulo4 and the last is4 modulo8. They are pairwise distinct and distinct from zero. Zero tests omit leaves. Thus each factor is nonzero with valuation<=0, and has residue1 whenever its valuation is zero, at both split places and at the inert place. This argument treats arbitrary negative test valuations too. [P06: trace/pole extraction] Write W for the product and ell=max_{v|p}v(W)<=0. The literal cleared trace equation w0^2-2w0*tau-2w1^2=0, with W=w0+sqrt2*w1, forces w0!=0 since W!=0, and tau=1/Tr(W^-1)!=0. The actual pole input is mathfrak_b=gamma^2 R/Tr(W^-1)^2, gamma=408/409, a p-unit. Trace gives v_p(Tr(W^-1))>=-ell, so v_p(mathfrak_b)<=nu+2ell. Its valuation is odd. The zero-preparation pole obstruction excludes negative odd valuations outside2,3,17 and hence forces it positive. Therefore ell>-nu/2. At a place attaining ell every factor valuation and every finite summand valuation exceeds -nu/2. Solving gives v(C2,C3,C4,G)>9nu/4, v(I)>9nu/8, v(C1)>3nu/8, v(D_H)>3nu/4, v(A2)>5nu/4, v(A1,A3,A4)>3nu/4 and v(B_i)>7nu/8. Integer rounding and even reciprocal valuations give cross/Lucas orders>=3, input order>=1, and every reciprocal depth>=2. No implication is extracted from a radical-one detector alone. [P07: exact integer consumers] Positive reciprocal valuations put H,U_i in the primary formal group and H-P,V_i in its conjugate formal group, with the identity convention at zero numerator points. Write H=[m]R0+T0, h=m/2; all consumed p are odd. The integral identities `[m]U_i=[2u_i]H` and `[m-2]V_i=[2(u_i-n_i)](H-P)` kill two-torsion. At least one of h,h-1 is a p-adic unit, so one full logarithmic ratio is integral. The positive C_i transfers integrality of the corresponding plain ratio to the other; formal logarithm units then give integrality of both full ratios. Only now is the relative error of order4 used as an absolute congruence. For i>1, L_i=n_i mod p^3; for i=1, L1=n=beta mod p. Consequently, for Delta=t^2-a(a-2)t*b+(a-2)^2 b^2-1, one has M_j^3 | Delta for each j, and M_j | J when J is the reduced numerator of n-beta. These are separate divisibilities, with no factor c and no requirement that M0 and M1 be coprime. Rational beta is p-integral whenever an input congruence is consumed. If J!=0, |J|<=2B exp(h(beta)); if Delta!=0, |Delta|<=14B^4. These are the exact objects used below. [P08: aggregate source application] The actual aggregate input is the **sum** theorem, not the odd-label product theorem. On E0:Y^2=X^3-X let P1,P2 be the pointed-isomorphism images of P and its conjugate. The independent squareclasses d,d' in K(i) provide an automorphism fixing P1 and negating P2 while fixing Z[i]; applying it to a torsion relation proves their Z[i]-independence modulo torsion. For arbitrary nonzero labels, at any common pole of summands in one block, the pole coefficient is lambda_i^-2 sum_j (n_i/n_j)^2, a nonzero positive rational multiple of lambda_i^-2. Cross-block common poles would contradict that independence. Restriction to a nonzero elliptic subgroup direction therefore excludes every positive-dimensional abelian translate through the actual tuple in a sum fiber. This covers negative, repeated and noncoprime labels with no parity condition. [P09: family/normalization join] Apply final published Von Buhren2017 Theorem1.1, printed308, to A=E0^8, parameter P^1, and the integral projective graph closure of F+=(d/2)sum X_i+(d'/2)sum X_i'. The fixed finite-rank group is generated by P1 in each first-block coordinate and P2 in each second-block coordinate. Its points lie in the theorem's C(Gamma,epsilon) by taking error point0. Embed E0 by O(3O), and its product by Segre. Restriction of degree-n plane forms to a smooth plane cubic is surjective for all n>=1 from 0->O(n-3)->O(n)->O_E(n)->0 and H^1(P^2,O(n-3))=0. Tensoring and the Segre multiplication map provide the source's embedding condition(1.1) for the product. P^1 uses O(1). The actual tuple avoids all coordinate poles since P1,P2 are nontorsion and labels nonzero. A translate in a projective graph fiber through this tuple would meet the function domain densely and make F+ constant there, contradicted by P08; graph boundary components do not change this. F+(tuple)=2T. The fixed product canonical height is a positive multiple of the sum of all eight coordinate canonical heights, hence twice a multiple of the four original heights. Since h(2T)<=h(T)+log2 and hhat([n_i]P)=n_i^2 hhat(P), hhat(P)>0, the theorem gives fixed M1>0,M2>=1 with B^2<=M1 h(T)+M2 for every tuple. The constants are fixed before x,N or labels; the additive M2 is ineffective, exactly as in the published source. No full rank-one theorem is required for this aggregate step. [P10: height domain and maps] Retain the sibling full-family supplier h(u)<=C2 M_S10(u)^alpha, alpha=2103/1024<7/3, on every rational argument avoiding its four finite contacts; the full geometric/modular/height chain supplying it remains a separately reviewed interface. Put B_Sigma=product of primes in Sigma outside S10. Squarefreeness gives M_S10<=B_Sigma M_Sigma, even at deep contacts, so h(u)<=H M_Sigma(u)^(7/3), H=C2 B_Sigma^alpha fixed. Apply this to u=2+1/t_j and u=2+1/(2-t_j). Their excluded finite contacts correspond exactly to the ten excluded t_j contacts, and t_j=0,2 are already excluded; the inverse projective row norms are3 and7. If h(t_j)>log7, multiply the two inequalities for h(t_j)-log7. Outside Sigma the two block radicals are disjoint and their product is exactly M_j, giving h(t_j)<=H2 M_j^(7/6), H2=log7+H. If h(t_j)<=log7 it follows directly, including t_j=1 with M_j=1. Thus the argument covers the full rational solution domain, not a bounded example. Because t0!=0 and t1=Tt0, h(T)<=h(t0)+h(t1). If M_j^k divides the same nonzero integer J0 separately for both j, then B^2 <= C |J0|^(7/(6k))+M2, C=2 M1 H2. This is the source-to-physical-radical join; no product divisibility between scales is asserted. [P11: a single soundness threshold] The actual scalar-S/four-contact argument gives c>0, c>=149^L_N and h(beta)<=C3*149^-L_N*c(1+log(c)/3), C3=2^330770, L_N=floor(9(2N-1)/4)+1. This uses c=|a|, not |a-1|. For Delta!=0, P07/P10 imply B^2<=D B^(14/9)+M2 where D=C*14^(7/18). Hence B<=K_D=max(sqrt(2M2),(2D)^(9/4)): when B^2>2M2, move M2 and divide by B^(14/9), obtaining B^(4/9)<2D. Choose N_D so 149^L_N>K_D for all N>=N_D. This is possible by monotonicity and excludes Delta!=0 uniformly in every x and every solution. [P12: Lucas lower bound and input threshold] With Delta=0 and a,t,b nonzero, set A0=|a|=c and w=sign(a)(a-2)b. Then t^2-c*t*w+w^2=1. The elementary norm-one unit reduction in Z[(c+sqrt(c^2-4))/2] gives |w|=U_j and |t| in {U_{j-1},U_{j+1}}, where U0=0,U1=1,U_{j+1}=cU_j-U_{j-1}. Nonzero w excludes j=0. Reduction modulo c-2 (if a>0) or c+2 (if a<0), together with |a-2| dividing w, gives j>=c-2. Induction gives U_j>=(c-1)^(j-1); hence for c>=8 the weaker bound log B>=(c-5)log(c-2) holds. If J!=0 and B^2>=2M2, P10 at k=1 gives log B <= K + epsilon_N c(1+log(c)/3), K=(6/5)log(2C)+(7/5)log2, epsilon_N=(7/5)C3*149^-L_N. Choose N_e so epsilon_N<=1 for every N>=N_e. Put K_c=max(exp(12),2max(K,0)+1,8,sqrt(2M2)+1). For c>=K_c, (c-5)log(c-2)-c-(c/3)log c >= c[(log c)/6-1-(log2)/2] > c/2 > K. Here c-5>=c/2 and c-2>=c/2 hold. Choose N_c so 149^L_N>=K_c for every N>=N_c. Then J!=0 is impossible. Thus n=beta is integral. Writing x=u/v reduced with v>0, integrality forces v^2|51; since51 is squarefree, v=1. The same fixed N_S=max(N_D,N_e,N_c) works for every rational input and all solutions at every N>=N_S. This construction is existential, and K_D,K_c depend on the ineffective M2. [P13: uniform scalar completeness threshold] For integer x, the retained four-factor scalar r4=beta(beta-1)(beta-2401)(32beta+343) is an integer>=1. With R_s=A_N r4/(1+149^2 r4), one has R_s>=A_N/(1+149^2), independently of x. Choose the fixed compact real neighborhoods from the scalar-S proof before finite progression moduli; on them Z0,Z1,F are bounded by constants M_Z,M_F independent of x,N. The actual parity-normalized scalar is W_s=Z0^4/R_s^10+Z1^2/R_s^5+F^2/R_s^4, b_s=R_s(R_s+9792)W_s^2. For R_s>=1, W_s<=K_W R_s^-4, K_W=M_Z^4+M_Z^2+M_F^2, so 0=N_C. This is uniform in the integer input. F!=0 and real positivity exclude b_s=0. The neighborhoods and constants are those of the original scalar functions, before the later root-label-dependent comparison neighborhoods; the latter need no bound uniform in labels. [P14: scalar finite preparation in correct order] For a particular x,N, put D_s=1+149^2 r4. Both gcd(A_N r4,D_s) and gcd(A_N r4+9792D_s,D_s) equal1; thus R_s(R_s+9792) has square reduced denominator D_s^2. Direct integer parity/congruences give v2(R_s) in {2,3} and v3(R_s)=v17(R_s)=0. Hence 1+9792/R_s is a square at2,3,17 (its difference from1 has depth>=3,>=1,>=1 respectively), so R_s(R_s+9792) is a square there. At its finite odd-valuation numerator support choose an initial H formal enough; then choose the positive scalar index a still deeper and arbitrarily large, with [a]P in the fixed real neighborhood of O. This makes the three terms of W_s integral at that support and deep enough at2,3,17 for the actual zero-preparation pole bounds. Compact-group recurrence in finite-index progressions gives the finite and real conditions together. After a is fixed, continuity preserves the scalar conditions for H in one progression h=h_* mod L, with64|h_*,64|L and the component-order divisibility included. This ordering permits support and depth to depend on x,N; the common N_C does not. The retained scalar/root chart guards hold on the chosen neighborhoods. No new condition a=1 mod any modulus is imposed. [P15: root labels, coefficients and incidences] Put n=beta. For the chosen positive a>=3 choose j>0 divisible by a-2; with the same Lucas recurrence put b=U_j/(a-2) and t=U_{j+1}. Then n,a,t,b are positive nonzero integers and Delta=0. Fix their four root points and T>0. Fix compact real neighborhoods for H and all comparison rational functions avoiding their denominators and chart guards. These neighborhoods may depend on the fixed labels. Enlarge the finite preparation set to contain Sigma, scale numerator/denominator primes, scalar-support primes and fixed guards. Choose D>0 so all numbers D^2 theta/a_* are integers for a_* in {1,T}, theta in Theta_2, and so v_p(a_*lambda)<0 with the needed depth at every prepared p when lambda=u/(D^2 w^2). Use coefficients c_{a_*,theta}=-D^2 theta/a_*; merge equal coefficients **across** detectors, retaining each incidence. Within a detector distinct theta give distinct coefficients; at least ten distinct coefficients remain. The zero contact supplies coefficient0 and the prime u. This is exactly the old center-two pattern. [P16: prime theorem quantifiers and CRT] Apply the actual V15 simultaneous-prime theorem to this fixed list, prescribed divisibility q|w, fixed avoidance modulus M=rad_odd(3*17*L) with7 omitted, and arbitrarily large slope/prime lower bounds. Choose these data before the analytic asymptotic parameter (call it X_sel, distinct from input x). Its fixed-field proof first fixes the cutoff so positive cutoff mass exceeds the full obstruction tail, then fixes the ray field, then lets X_sel grow along sufficiently slow diagonal strip widths. No uniform estimate in x,N,T,D,L or number of contacts is needed. The selected Gaussian generators have residues9 mod16,1+3i mod7,i modM; adjacent norms d_minus,d_plus have v2=6,2 and cross-sign gcd4, and all odd factors avoid L. Define U=lcm(d_minus/4), V=lcm(d_plus/4). Full two-torsion makes d/4 kill the reduction of the double P; gcd(U,V)=1, gcd(U,L)=16 and gcd(V,L)=1. Thus h=0 mod U, h=1 mod V, h=h_* mod L are compatible. The real progression lemma gives arbitrarily large h in the original fixed compact neighborhoods. Set H=[h]P, U_i=[hn_i]P, V_i=[(h-1)n_i]P. This gives natural formal depth>=1 at the correct primary/opposite places and preserves the scalar data. The optional strengthened separation p_i not dividing d_j is available from the weighted strip argument; natural depth1 alone suffices for this 214 construction. Every merged repeated coefficient carries one prime/generator and all its incidences; CRT is imposed once and supplies all those incidences. [P17: changed factors at the actual fresh witnesses] At any selected contact p, the slope and prime cutoff make p avoid D,w and both scale supports; selected primes are distinct and larger than w^2. Thus exactly its indicated contact numerator has depth1 in each incident detector, delta is a unit, v(q)=-1, every reciprocal test has depth>=2, and every C_i,I,G has depth>=4 by the full ratios n_i and relative formal error4. The changed factor terms have minimal depths q^5 C_i^2:3; q^5 I^4:11; q^3 A2^2:1; q^2 D_H^2 or q^2 A_i^2:2; q^2 C1^4:14; q^4 B_i^4:4. Therefore W=1 mod the selected place. At the opposite place the P05 residue argument applies without a formalness assumption: W^-1 reduces to0 or1, so Tr(W^-1) reduces to1 or2, a unit since p is odd. The new actual pole input mathfrak_b has valuation1 there. This is a fresh order-three-factor proof, not an appeal to the old order-four conclusion. [P18: all other primes and completions] For each physical scale a_*, the exact identity is R = [a_*^10 product_{theta in Theta_2}(u-(theta/a_*)D^2 w^2)] / [(D^2 w^2)^10 delta^10]. Every non-prime factor here is a rational square, including a_*^10 for a nonsquare scale; each contact coefficient occurs once within a detector. Hence outside its selected prime numerators, v_p(R) is even. This includes primes dividing delta. Multiplication by gamma^2 and division by the squared trace preserve that parity. At prepared p, D and the fixed divisibility q|w make R=product_theta(1+(1-theta)/delta) as close to1 as needed and the weight q=R^-1-1 arbitrarily small. All positive exponents in the changed factors therefore make test terms small there uniformly on the fixed p-adic neighborhoods, and Tr(W^-1) is close to2. Choose closeness so R is a square at2,3,17. As v2(gamma)=3 and v3(gamma)=v17(gamma)=1, the actual pole input is a local square with depths4,2,2. At both real embeddings, as lambda grows on the fixed neighborhoods, R=product_{r=1}^5(1-alpha_r/delta^2), alpha_r>0, so q>0 tends to0; all factors are positive and W tends to1. Thus mathfrak_b tends to gamma^2/4 in(0,17). The prime theorem permits an arbitrarily high fixed slope after these neighborhoods are chosen. This proves positivity and a nonzero trace globally, and q0,q1>0 so all new inverse weights exist. [P19: actual pole theorem and its curve domain] The input just constructed has square reduced denominator,00 squarefree,l=11 mod12. These curves have j=35152/9, are non-CM, and are distinct from the CM index curve E/K. RC is applied separately to these exact two parameters; no constants uniform in them are required. The published forward analytic-rank-one input then gives rank1 and finite Sha, and Kummer gives points with the required nonzero abscissa class. The zero-preparation proof explicitly rescales those two points to produce e,p,q,s for this actual new input. Distinct detector certificates have independent structural coordinates and can be chosen separately after both inputs are fixed. Zero auxiliary witnesses remain allowed, and the finite conic/support charts include those branches. The later CRT or prime choices are not revisited. [P20: final quantifiers and limitations] For every N>=N_C and every integer x, P14-P19 construct rational witnesses (existentially) in the original physical system, with the changed factors and nonzero inverse weights. Reconstruct deleted fourth powers and apply the exact raw224 coordinate maps and Laurent row identities; this gives a rational zero of the fixed 214 residuals. That argument allows the labels, selected primes, finite moduli, slope, final h and pole certificates to depend on the given (N,x); it imposes no further lower bound on N. Taking N0=max(N_S,N_C) combines this with P11-P12 for all larger N and every rational input, proving P01 relative to the explicitly retained interfaces. This author did **not** produce a full numerical 214 witness tuple, even for a sample input; no sample could establish universal completeness. This packet gives no numerical defining threshold, no Busy Beaver coefficient replacement, no independent acceptance of itself. Full Protocol D closure additionally requires the fresh eligible review and exact assembly of unchanged RC/analytic and height foundations, not a green structural check.