III. Heights and the square-root exponent
The arithmetic problem is to bound a weighted prime-power sum by its main term plus a controlled error. The proof compares the slope of the same row lattice with the slopes of a family of column lattices. At a fixed translate, the columns are counted with their actual ranks. The resulting inequalities can then be averaged with equal weights over translates. For the fixed-level extension we use the pointwise version before averaging over the subgroup. All source references are to the release at commit e7c8972c701d3396947808ab20cc16dc14623d33. Word and translate means · Mean and pointwise bounds · Fixed-level comparison
In rStation paper 192 v2, span_of_dilates derives the CM balance identity and explains how a hypothesized bound on the arithmetic gap yields the prime estimate at one host. Its Theorem 4.9 uses the gap bound, Proposition 4.12 gives the per-host implication, and Corollary 4.24 states the window conclusion with the host, compatibility and scale hypotheses of §4.6. The additional derivation begins in §1, where the finite-place indices give the adjusted column slopes. Section 2 compares their rank-weighted mean with the slope of the common dual row lattice through the integral period maps and the frame rank estimates. The signed height estimate cancels the common divisor term, and §3 uses the finite-prime loss bounds to obtain the same bracket. The comparison is conditional on fitted data at the same cutoff. Sections 4–6 use the shared arithmetic readout and square-root normalization, with a uniform mean over translates also available in the ordinary case under mean loss bounds. Paper 192 v2, Theorems 4.7 and 4.9 and Propositions 4.12–4.13 · Rank filtration · Gain comparison · Pointwise gap bound
The integrated refinement in §2 gives a linear spread term and a centered Gram error. With the general size bounds, §§3, 5 and 6 give a one-sided estimate of order in the ordinary and fixed-level cases, where . The original comparison also appears throughout.
Proposition (Integrated frame bound)
Fix one label with a finite nonempty word set . Over a totally complex number field , let and the common target be Hermitian lattices of positive -ranks and , respectively. Use the normalized slopes of §1, with the degree divided by rank and , the extremal Harder–Narasimhan slopes. Let be -linear maps of the ambient spaces with . For each take a real with whenever .
At a common embedding suppose there are matrices with , where projects onto the retained rows, is a single common invertible target change, and the are invertible source changes. Assume , , , and . For positive , set and assume with Hilbert–Schmidt norms in the frame coordinates.
Write for the Harder–Narasimhan slopes of minus , indexed by with each slope repeated according to its graded -rank. Put If satisfies , then This is (F), proved in §2. If is the fraction of adjusted column slopes strictly below , its integral on is the rank mean of , where . The exact mean makes the mean of the full deficits equal to the mean of the excesses , bounded by . This also gives zero integrated source deficit when .
1. The column and row lattices
Fix a finite nonempty set and, for each , a finite nonempty word set . A translate depends on . In the full translate calculation , a finite abelian group. In the fixed-level calculation there is a map whose image contains the required subgroup. Write for the number of lifts, for the set of tops, , for their sizes and . The common size parameters are . Here is the reference size, the reference thick weight, the thin scale, a size parameter for the lattices, and logarithmic bounds. Fix and positive weights for and . Scalar data · Lattice specialization · Translate coverage
At a fixed , the coefficient field is totally complex, and each lattice is a finitely generated -module spanning its ambient -space, with a positive Hermitian metric at every infinite place. Put . In the release's normalization the slope is , where is the negative log of the Euclidean covolume divided by the rational rank. Apply the same affine change to the largest and smallest Harder–Narasimhan slopes. The normalized degree is . For a line its degree is Classical: the determinant-line formula for degree (with the displayed normalization) and duality of Harder–Narasimhan slopes give and with dual metrics. Write for the common row lattice, realized isometrically in the retained row coordinates. Slope and map-height conventions · Duality and row covolume
For each word there is a full column lattice , its version after restriction at the type prime , and a sublattice of a fractional-ideal multiple of . The retained rank is strictly positive, and . For each finite place , set , let be the ramification index, and use . The finite set of contributing places contains every place over each rational prime it meets. Given finitely supported integral , set for either shift and put The use of powers with a negative exponent in fixes the direction of all the slope shifts. The maps are integral from to , and bounds the height for . Supplied by the geometric papers: the lattices, the integral maps, and the restriction and local estimates below are for one and the same family. Family and local hypotheses · Division and map height
Put , and . Let be the local gain, the common divisor term, and the type prime exponent term. The family supplies Here is the source's , and its . Both and depend on the translate alone. Write for the other finite-place charge, with , and . The identity follows from , with the residue degree. Local definitions · Arithmetic identification
Positive-rank comparison (proved here from the local inputs). A full-rank inclusion over a discrete valuation ring has determinant valuation equal to the length of the quotient. Applying this at every finite place, and using the ideal twist formula for degree, gives For the maximum slope, each sublattice of lies in . At a fixed rank its saturated enlargement in that lattice raises degree by the nonnegative index term, while the ideal twist moves every slope by . This proves the second line. The positive is also necessary in the first line. For the floor shift , the contribution to is exactly , including when . Length and slope conversion · Local gain identity · Conversion to place sums · Prior height calculation
Supplied by the geometric papers. For all words, , , , and . At one has and , with . Off these places, . The rank identity and the loss in passage from the full lattice are The estimates also give , , , and . Restrictions and length bounds · Size majorants
Indeed with , by Bernoulli's inequality. The part of at is at most . The other part is at most . With the preceding loss of slope this proves the following inequalities, where refer to , and to : The upper inequality holds already with zero in place of under the local hypotheses just used. We keep the released scalar allowance. Positive-rank scalar comparison · Transfer of the bounds · Word loss
The selected family packages the same calculation as a place-sum input. In that notation the length contribution at is supplied directly with bound , the nonnegative rounding charge is at most , and the weighted sum of the other local allowances before dividing by the retained share is . The bound , the restricted-slope comparison, and the allowance of give exactly the same by the preceding argument. The geometric supply for this route can give these place-sum hypotheses directly. Selected place estimates · Selected scalar bounds
2. The slope bound and the main term
The constants are nonnegative. Write for Euler's constant and for the centering constant of the first sign character defined in §4. For positive , set , and set , continuously at zero. Define The number is the sum of the logarithms of the two comparison primes in the constructed family. These are common across words, whereas is the prime used for the restriction of word . The live constants set , with the values of furnished by their respective estimates. Main and losses · Prime and count formulas · Choice of constants
Supplied by the geometric papers. At every translate and for every word, the following inequalities hold. In the pairwise inequalities both words belong to the same . The first inequality is a bound for the dual row slope and the map height. In the divisor assembly the upper bound for the base slope contains in the notation of that assembly. The comparison subtracts on the left, and the scalar is precisely that . The same finite sum entered the column comparison with . At the divisor stage the term is the type-prime part of the height. The other top contributions are obtained from once per lift. The geometric papers owe the estimates of the local indices, the Gram determinants, and the archimedean and factorial terms that give these bounds. Divisor assembly · Normalization and transfer · Height and full-lattice slope bounds
The algebra of that first inequality can be seen before the error is absorbed. In the assembly write the height as a type-prime term plus the lift heights, less the archimedean main term and the covariance normalization. The latter normalization cancels the same term in the upper bound for the base dual slope. For each lift and position , the lower estimate of the archimedean term is subtracted from the lift height's factorial term, leaving at most plus its error. The degree multiplying is in the assembly, where is its degree parameter. The corresponding base slope term is . Their difference leaves and costs at most . Thus the exact remaining majorant there is the sum of the base-to-row and base errors, the lift-height errors, the archimedean and type-prime errors, the factorial comparison errors, and those two degree discrepancies. The geometric input bounds that sum by . Assembly with each error · Selected height and count inputs · Prior reading of the signed height
Supplied by the geometric papers. At each translate take matrices and positive scalars . The projections of to the retained rows are the matrices of at a common embedding in fixed invertible changes of coordinates. Write for the dimension of this retained space, with . For the frame conditions are The Hilbert–Schmidt norm is the square root of the sum of squared matrix entries in orthonormal bases. The matrix identity is , with retained projection , invertible source changes , and a single invertible target change common to the words. These changes can be nonisometric. In particular . The parameters lie in , , , respectively, with and . Frame operators · Realization and inequalities
Define the rank mean and the spread Then . The geometric height input also gives with . For , let be the fraction of all column slope dimensions with adjusted slope and the fraction of retained row slope dimensions with slope . These fractions can be taken in -dimensions because the same factor converts them to rational ranks. Live mean and spread · Scalar height window
Proved here. The high Harder–Narasimhan step of at maps into the high step of at . Indeed the minimum slope of the domain step is at least , and the maximum slope of the quotient of by its high step is strictly less than . A nonzero composite would give a positive-rank quotient of the domain step mapping injectively to a saturated sublattice of the target quotient. The two induced metrics make the composite's squared-norm height at most , since the subspace uses the restricted metric and the quotient uses the minimum norm of a lift. The determinant inequality bounds the slope of the domain quotient above by the slope of the saturated image plus , contradicting the two slope bounds. This uses the classical Harder–Narasimhan characterization of extremal slopes by saturated sublattices and quotients. The inclusion holds also when the high step or the composite map is zero. Slope and height conventions · Threshold inclusion in the release
At the chosen embedding let be the orthogonal projection onto the transported high column step. Put , , and . Cauchy–Schwarz and the Gram bound give . In fact the trace pairing of the Gram deviations with the identities has modulus at most . Pairing instead with bounds the discarded trace by . Thus the discarded trace fraction is at most . Since , . The trace Cauchy–Schwarz inequality gives . The image of is the span of the high column images. Its codimension cannot increase under projection to the retained rows. The image inclusion just proved then gives the dimension estimate Here the bound on the ambient codimension fraction uses . Threshold comparison
Proved here. The rank mean of the adjusted slopes is , and every adjusted slope is at most . For the mean is at most , so . Since every row deficit below is at most , . Integrating the dimension bound gives a loss at most The first term is zero at , when every adjusted slope is . This implies the source's scalar comparison, which also allows as a nonnegative loss. Let , and suppose . The inequality applied to gives The last step uses and . The same bound holds at . Scalar frame comparison · Logarithmic estimate
Integrated refinement of the frame bound
The “Pointwise bracket” section of Error powers for the same ordinary family integrates the actual source deficits. The “Variance and edges” section of Frame energy and the HN slope bound centers the Gram operator at its actual mean trace. The proof of (F) uses both ideas at the filtration thresholds. The high step at a normalized threshold uses the last grade of slope at least , with index zero if no grade qualifies. The corresponding Euclidean threshold is , with . Thus ties belong to the high step and both low counts are strict. Pointwise bracket · Variance and edges · Rank bound in the release
Fix . Use , , and . Put , , , and . Expanding the square and using gives The blocks have total squared Hilbert–Schmidt norm . Their trace pairing with the centered Gram blocks is , since the sum of the centered Gram traces is zero. Divide Cauchy–Schwarz by and use to get , where .
Let project orthogonally onto the complement of the high frame span. Since is supported on that span and , we have . The trace-zero operator has squared Hilbert–Schmidt norm at most . Pairing it with , of squared norm , shows that . The high column spaces are transported by before forming their orthogonal projectors. Their images after row projection lie in the high step of transported by the common . Scalar extension preserves rank, and row projection reduces the dimension by at most . Hence . These dimension comparisons hold for the given invertible basis changes.
Use and to obtain . An adjusted column slope contributes to the integral of the strict low indicator. The rank mean of equals that of , which is at most . The integral of is the mean positive row deficit, since . Integrating gives This includes zero spread and empty steps.
Subtract from the lower column bound. The divisor inequality gives , which cancels the common divisor term. Apply the lower bound for the full-column mean, the upper bound , and , with . For the main term use the exact identity In the second line cancels the type-prime term, and the two terms involving cancel each other. Averaging over words with the positive weights gives Main-term identity and lower mean
To bound , subtract the lower bound for from the upper bound for before taking the mean in . Both and cancel. The variation of is bounded by by subtracting and adding the common center and the two terms . Thus The maximum over leaves the estimate unchanged because the right side after the mean in is independent of . Variation and spread · Count bound for the mean loss
Put , and . Substitute the bound for into the row comparison and use the lower estimate for . The coefficient of is one on each side. This gives, for each , Gain comparison
Let be the prime-power sum for a translate, defined in §4. Supplied by the geometric papers: the finite-prime loss bounds control and each by , pointwise for the translates used in the fixed-level route. For the mean theorem it suffices that these same bounds hold on their two uniform means over a nonempty set . Put and . Adding and using the nonnegative coefficient of gives the pointwise inequality Under mean loss hypotheses the same computation gives , where is the displayed expression with in place of . This is the uniform translate mean when . In the fixed-level route the comparison is pointwise on , and each subgroup translate is assigned a preimage in before averaging. Its weight in that average is independent of the number of preimages and the ranks of their words. Loss hypotheses and errors · Uniform mean · Pointwise consumer
3. Reducing the height error to a size bracket
Supplied by the geometric papers. For the same data, the size conditions used here are , , , , , , , , , , , , and . Also , for each , , and Here dominates the seven coefficients . The trace and Gram errors in must be controlled after multiplication by . The concrete scales are , , , , and , with . The estimates below use the displayed size inequalities and permit any scales satisfying them. Error windows · Concrete scales
For the integrated refinement put . It suffices to assume for this route. The original budget implies it since for . From (F) and the lower bound for , the gap at a label is at most The nonnegative coefficient of and its bound by allow the two loss hypotheses to be applied pointwise or after uniform averaging. In both regimes the resulting bound is .
Proved here. In this section write . Take with , so and . Put , , , and . Then , , , and . Direct substitution gives In the bound use . In the last bound the contribution of is bounded by . For the integrated refinement, , , and . The four coefficients of are bounded respectively by , , , and . It follows in both loss regimes that the gap is at most
For the original comparison also impose , with the original budget . Then . The same estimates give , and . The remaining terms of the pointwise error are bounded by , by the four inequalities and . Absorbing their coefficients with proves both original error bounds The hypotheses of this estimate are uniform over words and translates. The threshold depends on the size constants and . Absorption in the release · Explicit coefficient inequality
Let for the set of thick positions and . Supplied by the geometric papers: the upper bounds for , on the same lattice and prime data, are If , an upper bound on either arithmetic gap is consequently This substitution uses the identification of tops, weights, centering constants and with their arithmetic versions. It is valid pointwise on any translates covered by the construction. Tied data and budgets · Bracket and mean identification · Pointwise substitution
4. From the arithmetic gap to the character sum
Let be finite abelian, , and . Write , so . Put , , and Translation by identifies with . In Paper II's ordinary coordinates and . A position is thick at if for some . Coordinates · First sign and partner · Second sign and signed sum
Supplied by Paper II and the host construction. For each prime there is an inertia subgroup and a lift . The split exceptional set contains and excludes the distinguished prime . For primes in that set both and lie in . At , . At every other prime inertia is trivial. For a unitary character of , write for the mean of over , extended multiplicatively. For the principal character this function is identically one. Let be the primitive logarithmic derivative at for , and use the centered sums Here ranges over positive integers, is the von Mangoldt function, and is the ordinary notation used in Paper II, independent of the translate . Host inertia · Primitive coefficients · Ordinary centered sum
For a top , , and weight , define Here , and the unramified sum excludes the split exceptional set, whose contribution vanishes. We assume for every , so terms with vanish. At translate , the sum over of these 's is for , and the sum of the real parts of the corresponding 's is . Denote this gap by . Prime-power weights and centering · Identification
Proved here. Character orthogonality gives The complementary cosets are empty at split primes and have exactly one term at for each . This proves the prime weights in from Fourier inversion, including the coefficients and . For put . In the Fourier expansion of use and replace each resulting prime sum by its centered version. The principal character contributes the correction . Each other character contributes before summing over . The mean of on the complement of is for , for the principal character and for every other character. The centering corrections therefore add up to , including its . It follows that Prime-power expansion · Centering identity
Supplied by the host and weight construction. For every , depth and translate used in the average, put . Assume . The weight is positive, lies in at thick positions, and is at most elsewhere. The conditions also give the cutoff above, for all and , and the two cost bounds The weight tolerance and the row frame error occur in different estimates. Node and cutoff hypotheses · Thin and cost bounds · Fixed-level data
Classical input and elementary bounds. Use finite abelian character orthogonality and the identity . The elementary Chebyshev bound follows by comparing with for integer and summing at powers of two. Indeed each prime power in contributes one unit to the valuation of the binomial coefficient and all other contributions are nonnegative. For , summing the divisor identity gives . Bounding the floors and using gives The lower bound uses . Differentiate the continuous, piecewise smooth ramps between their breakpoints. Since , one obtains for and Also has derivative in of modulus at most . For the derivative is the negative of the corresponding twisted ; for the principal character it is . Integration between two positive weights gives These bounds give a choice of the absolute constant in the movement estimates. Chebyshev and weight comparison · Movement bounds in the release
Proved here. For every thick partner replace by , at cost at most per character and position. For a thin partner , so discarding its term costs at most . There are positions and characters with a prefactor . Thus the modulus of the error in the Fourier sum, for each , is at most Keeping the two position counts gives the sharper node bound , as used in Paper II. The coarser allowance gives the same released exponent. The thick sum left over is independent of . Write . The thick partners have for , so this sum is Single-node comparison and constant
Let with . Denote the characters trivial on by . Summing the character in over gives zero unless . For each surviving the factor is , since and the first bit is of order two. Hence This gives the in the normalizer Whenever , the modulus estimate for the nodes gives The fixed-level hypothesis of pointwise control implies the mean hypothesis in this display. Subgroup orthogonality · Subgroup comparison
In the ordinary case , the four survivors are , with signs at . Suppose the three nonprincipal survivors obey and with , as in the weaker form of Paper II's bounds. The movement from to contributes at most Before using the upper gap inequality, the Fourier calculation gives the two-sided comparison used for the ordinary normalization in Paper II: With fixed, eventually. In particular . Four-character signs and explicit bound · Ordinary translate identification
5. Division by the normalizer
Supplied by the geometric papers. At the chosen cutoff let and assume the additional size relations Then and . Proved here: for any real , if division gives the exact normalized bound In the ordinary case put . With the preceding size relations, the four terms of satisfy The last inequality uses . Thus . Since and , the normalized bound implies The two size costs and enter before density and logarithmic allowances. At the trial scale their exponents are and , with maximum minimized at . The fitted range for allows shifts of order in these powers. Size hypotheses · Size bound
For the integrated refinement substitute for in the same division. The four bounds on apply to the two parts of and to through the budgets in §3. With , , and , this gives For and , we obtain , including the other two terms in the division. In the ordinary case this holds for once and , where is defined in §4.
6. The fixed-level window input
The release also needs a bound for one fixed family of primitive characters, even when the Artin group varies with . Use the subgroup from §4, of index fixed by the level . Let be the kept characters. Supplied by the host comparison: contains the principal character, its characters have , and . Put and if is trivial on , otherwise. For each kept character nontrivial on , assume The host comparison gives for nonprincipal kept characters, and for discarded characters in . The primitive readout identifies the kept sum with As the host changes this is the same function, with the primitive characters and weights fixed by . In particular at the conductor-one level of Paper II. Host character and share hypotheses · Reading the fixed-level sum · Bound for discarded characters
Proved here. Let . For any nonprincipal character the movement bound of §4 and give . For a kept character , first move to at each . The remaining sum of phases is exactly if is trivial on . In the other case the deviation from is bounded by the share hypothesis. For a discarded character the real part is bounded below by less the movement error. For the principal character the movement costs at most . Thus the sum on the left of the subgroup comparison in §4, before its factor , is at least where one can take For fixed , eventually. Apply §5 with . For , put with . Then on increasing the threshold so . In the ordinary case the direct estimate in §5 gives the stronger exponent under Paper II's three-character bounds. One-character comparison · Size and discarded-character absorption
For the integrated refinement use the pointwise bound (G) at each subgroup translate in §4. With , the integrated estimate of §5 applies to . Adding gives . The cutoff conditions hold beyond one threshold for fixed and the uniform host constants, since .
Handoff to Paper II and the analytic window argument. In the notation of Paper II, , , and its is our . The in its error bracket is the reference size . The arithmetic input for its ordinary prime comparison is for the host used in that comparison. The pointwise form gives inequality (9) of Paper II. For the release's fixed-level route use for every . Fix a level , a window exponent , and . For every sufficiently small and every requested lower cutoff there must be an interval starting above that cutoff. At every real in that interval one has compatible host and geometric data with bounds uniform in as above. A different host and lattice family may be selected at each , with the same . The release uses along its geometric route, giving the one-sided bound on late intervals. For any choose so . Taking the lower cutoff beyond the thresholds in §§3–6 then gives throughout arbitrarily late intervals. That is the hypothesis of the analytic window theorem for the positively weighted family of primitive characters. Paper I proves the analytic implication. The existence of compatible windows is the supply hypothesis in §7. Paper I · Paper II · Window hypothesis · Host supply · Released exponents · One-point selection · Analytic consumer
7. The geometric inputs to be established in prose
The same window quantifiers in §6 apply to the integrated estimate with either the old frame budget or the allowance . The exponent bounds and the window exponent are fixed first. For each sufficiently small fixed , the size and host constants are uniform before requesting the lower cutoff . For either ramp let be its constant from §5 or §6. For any , take beyond the relevant movement and size thresholds and . This gives a unit-coefficient upper bound with exponent at every real cutoff in the supplied interval. Given , first choose small enough that . For the general size bound has .
| Implication proved | Supply needed |
|---|---|
| Local indices give adjusted column slope bounds (§1). | Full and restricted lattices with the stated ranks and metrics, and length, twist and gain estimates on the actual sublattices. |
| Integral comparison and centered integration give (F) (§2). | Integral period maps and height allowances, common-embedding frame realization, row retention and Gram and trace bounds, and the row minimum-slope window. |
| Signed cancellation and size bounds give the gap bound (G) (§§2–3). | The signed height estimate with the same , type-prime counts and slope bounds, the frame budget, and both finite-prime loss bounds, in uniform mean for the ordinary route or pointwise on labels covering for the fixed-level route. |
| The gap gives the prime ramp bounds (§§4–6). | The same host labels, tops and weights in the arithmetic identity, thick and thin estimates, size and character bounds, and, at fixed level, subgroup coverage, character shares and a fixed primitive sum. |
| Supplied prime bounds give the window implication via Paper I, also at discrete cutoffs for ordinary zeta (§7). | Compatible data throughout arbitrarily late power intervals, with uniform constants fixed for each before choosing windows. For ordinary zeta, the §7 corollary also permits data at cutoffs with its logarithmic covering bound. |
For the ordinary ramp or a fixed level , supply the windows just stated for each sufficiently small . At every sufficiently large in those windows, choose compatible data to meet all four requirements below, with the constants of §3 fixed on the family for that . The analytic estimates at the host must apply to the same Artin data, weights and tops. Tying the fields and data · Geometric input assembly · Fixed-level interface
Columns, periods and lengths. Construct the row and column lattices, their metrics and the rational period maps. Prove that the common row is the metric dual of , that the maps on the shifted columns are integral, that the full and retained ranks are positive with the stated restriction, and that every local length, shift and gain satisfies §1. In particular the finite index loss comes from the same sublattice used to define . The common divisor term used there must be the one in the height inequality. Place data · Period-integral lattice and shifts
Frame, counts and height. Establish the single-place matrix realization with positive normalizing weights, the trace and Gram estimates with the indicated row and column dimensions, and the height bound for each nonzero map. Prove the quantitative row minimum-slope window, the count and full-column slope estimates, the type-prime compensation and the corrected-height variation in §2. In the divisor estimate use the signed paired local determinant expression of the row covolume and bound its finite and infinite terms to obtain with . This step requires the common-divisor estimate for the selected family, in addition to the formal cancellation of . Frame hypotheses · Finite determinant expression · Global height input
Finite-prime loss. For each translate used in the pointwise argument, establish both and with . This entails the estimates on the orbits served at eligible primes, on the net gain , and on the omitted contributions after weighting by . It includes the distinguished ramified prime and the length charge . The two bounds on uniform means over suffice for the ordinary mean route in §4. Local gain and served orbits · Pointwise loss interface · Released loss exponent
Uniform supply and compatibility. For all cutoffs in the windows, prove the coverage of the required translates, the shared top and weight identifications, the size inequalities of §§3 and 5, and the thick and thin conditions of §4. In the fixed-level case include the share and primitive readout hypotheses of §6. In particular a comparison for one abstract choice of numerical slopes gives the scalar implication, and the needed existence statement requires slopes of the actual lattices built from the chosen host. Pointwise data and coverage · Compatibility equations · Quantified family supply
The geometric statements are inputs here. The height and scalar estimates above show how those statements would yield the ordinary and fixed-level prime comparison used in the series.
Ordinary zeta at discrete cutoffs
The ordinary-zeta conclusion also follows when the geometric construction supplies data at a discrete set of cutoffs in each interval. This consequence uses the explicit formula and the window theorem of Paper I. Paper I, Theorem 2 and its proof
Corollary. Fix and . Suppose there are constants such that, for arbitrarily large , a set has both of the following properties. Every has some with , and for every . Then every nontrivial zero of has real part at most .
Proof. Write over the nontrivial zeros. Paper I gives, with multiplicities and an absolute remainder, The classical zero count gives by dyadic summation. For real with , integration of and the bound on give . Thus for , At a sufficiently late supplied interval choose the node for each and take . The node bound gives throughout that interval. If , choose strictly between and . On arbitrarily late windows the maximum of is then at most . Paper I proves that these maxima tend to as . This contradiction proves the corollary.
For each fixed positive integer , one may take all in the window. Consecutive powers satisfy . At either end of a sufficiently late window the first or last power inside it gives the same bound. For the ordinary implication one may therefore impose the common geometric hypotheses at those powers only. For each fixed the covering constants and the positive may depend on , provided they are fixed before choosing arbitrarily late windows. The bound with at the nodes gives whenever . Taking to zero and using the functional equation gives RH under this discrete supply hypothesis. In particular, it suffices to take along integers and use at that exponent. The two scales and are then integers. The power is fixed while the starts of the windows tend to infinity.
One can also choose the two even ranks exactly. Fix below the common existence bound and use the nodes . At these cutoffs the lower endpoints of the two rank intervals are the even integers For a frame of order with , these choices satisfy , and for large , by comparison of the displayed powers of . Thus the rank construction applies to these very integers. The logarithmic mesh is , so the corollary applies on the same power windows. The auxiliary rank formula still contains a ceiling, the degree of the dyadic field must be a power of two, and the period and weight hypotheses must hold together on the chosen host. Construction of the ranks on a given frame