WorksMathematics
An asymptotic resolution of the Gao–Thangadurai conjecture on the Harborth constant
Abstract
The Harborth constant of a finite abelian group of exponent is the least integer such that every subset of of size at least contains elements whose sum is zero. Let be a cyclic group of order . Gao and Thangadurai \cite{gt} in 2004 conjectured that for all integers . We prove this conjecture for all sufficiently large integers .
Introduction
The Erdős–Ginzburg–Ziv theorem [4] states that for every positive integer , any sequence of integers contains terms whose sum is divisible by . Geometrically, these integers are points on the line whose centroid is again an integer.
Kemnitz [17] asked the corresponding question in the plane. How many points in suffice to guarantee that some of them have a centroid in ? The centroid of points is a lattice point precisely when both coordinate sums are divisible by . Reducing both coordinates modulo therefore gives an equivalent additive problem in , where denotes the additive group of integers modulo . Each residue class has infinitely many lattice representatives, so distinct lattice points may give repeated terms in . In these terms, Kemnitz asked how many elements of , with repetitions allowed, suffice to guarantee that some of them sum to zero.
Taking copies of each vertex of the unit square shows that terms do not suffice. Kemnitz [17] conjectured that the exact value is . This conjecture attracted considerable attention, with contributions on upper bounds and special cases [1, 5, 6, 22, 25, 26]. Alon and Dubiner [1, 2] proved that terms suffice for every . Rónyai [22] reduced the upper bound to for prime , and Gao [6] extended this bound to prime powers. Thangadurai [26] showed that terms suffice if some element occurs at least times. Gao and Geroldinger [7 Main Lemma 5.6] showed that the conjecture would follow from a lower bound on the number of affine hyperplanes avoiding the origin needed to cover certain sets over prime fields. Reiher [20] finally proved the conjecture for every positive integer .
The corresponding problem for sets predates Kemnitz’s conjecture. In 1973 Harborth [15] introduced the constant for subsets of . In the plane, how many lattice points with pairwise distinct residue classes modulo suffice to guarantee that some of them have a lattice centroid? In this case the sequence in is squarefree, so we may regard it as a set. For a finite abelian group of exponent , Gao and Thangadurai [10 Definition 1] define as the least positive integer such that every subset of of size at least contains an -element subset with sum zero. Following Marchan, Ordaz, Ramos and Schmid [18], we call the Harborth constant of .
The one-dimensional case is trivial. In dimension two, for , Kemnitz [17] proved that for every . Gao and Thangadurai [10] conjectured that these lower bounds are the precise values of .
Conjecture (Gao–Thangadurai, 2004 [10 Conjecture 1]). For every integer ,
The formula was already known for : goes back to Harborth [15], while Kemnitz [16] in 1982 established . Gao and Thangadurai [10] proved it for every prime and also established . Gao, Geroldinger and Schmid [9 Theorem 5.1] extended the prime result to and described the extremal sets of size . Guillot, Marchan, Ordaz, Schmid and Zerdoum [14 Proposition 4.1] established the next even case, . In a recent preprint, Graham [12] reports that for every prime . Together with the known cases , this would settle the conjecture for every odd prime.
In this paper we prove the Gao–Thangadurai conjecture for all sufficiently large integers.
Theorem 1.1. For every sufficiently large integer ,
Notation and preliminary results
Throughout, is a finite abelian group with additive notation and identity . Let and . For an integer , let . The exponent is the least positive such that . For , let be the quotient map and let denote the index. If , then .
We follow the sequence notation of Gao and Geroldinger [8 Section 2]. Let denote the collection of all finite unordered sequences of elements of , with repetitions allowed. For , its length is and its sum is . The notation means that is obtained by choosing some of the positions of . The sums of all subsequences of of length form the set
with . A zero-sum subsequence of length exists precisely when . A sequence is zero-sum free if none of its nonempty subsequences has sum zero. It is squarefree if its terms are pairwise distinct. We identify a finite set with the squarefree sequence of its elements. When we project that sequence into , we retain one term for each original point. Thus two equal terms in the quotient can come from distinct points of .
For nonempty sets , let and . In an expression such as , the summands are chosen independently from and may repeat. In , the summands are distinct. The stabilizer of is the subgroup . We call aperiodic if . In every subgroup is for a unique divisor . Its index is , and its cosets are the residue classes modulo .
A primitive row of is a coset of a cyclic direct summand of order . In coordinates adapted to that summand, the parallel rows are the sets . We call the height and the first coordinate the horizontal coordinate. Translating all points by preserves every -point sum, since it changes that sum by .
A coset progression with directions is a set of the form
where , , and the are positive integers. It is proper if each of its elements has a unique representation with and . When , we call a proper progression a box. Its volume is the number of parameter choices, .
The following sumset bound will control stabilizer indices. If with nonempty finite sets and , Kneser’s theorem [3] gives
We also need zero-sum subsequences of prescribed length in quotients. By the Erdős–Ginzburg–Ziv theorem (EGZ), terms in suffice for a zero sum of length [4]. By Reiher, terms suffice in [20]. A quotient of with exponent embeds in by the invariant factor theorem. Consequently, for every positive multiple of ,
Indeed, successively remove disjoint zero-sum subsequences of length . Before the last choice at least terms remain, so Reiher’s theorem applies at every step. Their union has length and sum zero.
For restricted sumsets in a cyclic group, we use the following bound. If is a set of elements of a cyclic group, then for [11 Theorem 1.1]. For the sumset is , and for it is . The exception in the general abelian-group theorem requires a coset of an elementary 2-group of size at least four, which cannot occur in a cyclic group.
The remaining two results describe long sequences in cyclic groups. Savchev and Chen proved that a zero-sum free sequence in of length has, with respect to a suitable generator, positive representatives whose sum is [23]. They also proved that if and a sequence of length in has no zero sum of length , two values have combined multiplicity at least [24 Corollary 7].
Proof of the main theorem
By Kemnitz’s lower bound , it suffices to prove is also the upper bound for all sufficiently large . Throughout this section, a counterexample is a set with and .
We begin by counting representations of a fixed sum by tuples of points of . The next lemma bounds their total number when no individual point occurs too often among the representations of any one sum.
Lemma 3.1. Let , . Consider a family of ordered -tuples of distinct elements of , where . Suppose that at most tuples with a given sum contain a given element of , with . The family has fewer than members.
Proof. We show that failure of the bound produces an -term zero-sum subsequence of . Let count the tuples with sum , choose with maximal , and set . For let . If the asserted bound fails, then
Also by incidence counting, so . For , let , , and . Kneser gives
Hence . The set contains . Round up to a multiple of , so . Since , we have . Adjoining copies of the singleton therefore produces a sumset containing $H. Every such sum can be realized with distinct points of . Choose a tuple for each prescribed layer value in increasing order of . Earlier tuples use at most points, and hence exclude fewer than representations at layer . At appended positions they exclude fewer than , and the value has at least representations. Since and divides both and , the integer is positive and divisible by . Outside are at least points. This is at least , since and . By (1), choose of them with sum in , and cancel that sum with the points inside . □
Lemma 3.2. Let . For of size , one has . Also, if are disjoint of sizes , with , then
Proof. We first prove the energy bound. Let count ordered pairs from with sum , and set . If the pairs with account for at most pairs, at least ordered triples have distinct entries and all three pair sums of multiplicity at most . Fixing a sum and one element leaves at most such triples. Lemma 3.1 gives strictly fewer than . Thus the pairs with contribute more than to .
To bound the intersection, consider tuples with and , and unequal entries within each pair. Since , all four entries are distinct. There are at least such tuples. Given a sum and an element of , choose its position and . The last pair has at most choices. The same bound holds after exchanging and . Apply Lemma 3.1 with , and ; the hypothesis follows from . □
Lemma 3.3. Let , and . If , a counterexample contains points in a box
with unique representations, where , , and , with .
Proof. We first find a subset with small doubling. Take of size . Since , Lemma 3.2 gives , where . The energy form of the Balog–Szemerédi–Gowers theorem, in the explicit version of Reiher and Schoen [21 Theorem 1.2], with its parameter , gives satisfying
The Plünnecke–Ruzsa inequalities [19], applied with base set and summand , give . Hence .
To obtain the required box, we apply the result of Green and Ruzsa [13 Section 5, (5.1)], which gives a proper coset progression of rank at most and size at least . Averaging the identity
over gives a translate containing at least points of . This number is at least , since
The subgroup part of has at most two invariant factors. Use one box coordinate for each such cyclic factor and translate the origin of each parameter interval to zero. Properness gives unique representations, so , each , and . Take the points of in this translate as the points of the lemma. □
The box in Lemma 3.3 may have only one long direction. To handle this possibility we bound the number of points of a counterexample on a primitive row. We start with a row containing more than four fifths of its points.
Lemma 3.4. If , every primitive row contains at most points of a counterexample.
Proof. We first show that a row of this size attains every horizontal sum for every length between 2 and two less than its size. Suppose a row contains more than points. Choose a largest row in the same direction and translate it to height zero. Let be its set of first coordinates and . We claim that whenever . By complementation take . For even choose two points summing to the target. For odd , first fix one point, then choose two others for the adjusted target. This works as after removing at most one point, and at most two solutions have equal partners. There are initially at least disjoint opposite pairs in . Even after losing three pairs there are at least left, since . They supply the remaining points in pairs.
We next exclude . If is odd, the full row sums to zero. If and is even, take outside heights. These heights have a nonempty zero-sum subsequence of length less than . Otherwise every -subset would itself have sum zero by the partial-sum argument in . Comparing two such subsets differing in one entry makes all heights equal, exceeding row capacity. The full row also covers every target at sizes , so one can cancel its horizontal sum with the required number of row points.
For , we use the outside heights to complete a zero-sum selection on the chosen row. Let and take outside points, available in either parity. A zero-sum subsequence of their heights of length in would leave between 2 and places on the chosen row, where any required horizontal sum is attainable. Suppose there is no such subsequence. Some generator occurs at least times among these heights. To see this, if there is a minimal nonempty zero-sum subsequence of length or , delete one term and apply the zero-sum free sequence theorem of Savchev–Chen. Its remaining positive representatives have sum at most . Restoring the deleted term with its representative in gives a positive multiple of strictly below , hence total and at least ones. Otherwise remove disjoint minimal nonempty zero-sum subsequences until none remain. Each has length at most . Their total cannot first exceed , since it would be at most . At least zero-sum free terms remain. Their positive representatives have sum at most , so at least are ones.
We now bound the multiplicities of the remaining heights. Any with $3 \le a \le n - r - 1$ would give a forbidden zero-sum subsequence with copies of . Two copies of $2g$ would do the same with copies of . The other possibilities are , . Their weights total at most , since a first prefix exceeding could be completed with copies of to a zero-sum subsequence of length between and . Such a prefix contains at least two negative-height terms and needs at most copies of , which are available. The height occurs at most times, since we started with a largest row in this direction. Thus for the outside count is at most . For , if occurs take one point of each height , and from the center. The -row choice and center choice give respectively at least and horizontal sums, so their sum fills . Here we use the fact that whenever and . If is absent, the outside heights are copies of and one . Take one point of height , of height and one from the center. The latter two choices each give sums.
Lemma 3.5. Let and . After deleting at most points of a counterexample, one can select remaining points with height sum zero in and first coordinate sum zero modulo .
Proof. We first prove the assertion for . Let be the sequence of remaining heights. If , cyclic EGZ gives the required selection. Otherwise, for , we have and . If the selection fails, Savchev–Chen gives two heights with combined multiplicity at least , contrary to Lemma 3.4.
For , we seek disjoint blocks whose union gives the required selection. Let be the number deleted. A block consists of points whose two coordinate sums vanish modulo . Its value in is its height sum divided by . Try to choose disjoint triples of blocks with three different values in each triple.
If such triples exist, we obtain the selection from a maximal collection of disjoint blocks containing them. Reiher leaves at most points, so the number of blocks is at least . We need blocks of value sum zero. EGZ suffices if . Otherwise , since and . If there were no such choice, Savchev–Chen would force all but blocks on two values, contradicting the triples.
If fewer than triples can be chosen, we concentrate the remaining heights in at most two residue classes. Discard a maximal collection of such triples. Among the remaining points take a maximal collection of disjoint blocks with pairwise distinct values and discard them as well. There are at most two such blocks, so the total number discarded is less than . Every further block has value in the same set of at most two values. All but at most remaining heights belong to modulo . Indeed on the real circle of circumference mark every lift of . Each of at most open arcs contains fewer than points counted with multiplicity. Otherwise Reiher supplies a block in that arc. Its average height, computed in an open interval lifting the arc, is an integer strictly between the endpoints and represents its value modulo , which is impossible. If is empty there are at most remaining points by Reiher’s theorem, again within the stated bound.
We next reduce the remaining points to at most two families with sufficiently many copies of each label. For each surviving height residue modulo , use a fixed integer representative and express the height as . Label the points in this family by . Discard all label classes with fewer than elements. There are at most classes, so this discards fewer than points. At least points survive, since for . If one family has at least points, group equal labels into -tuples. Fewer than points are left over. Any of the groups have both required sums, since the heights are with total divisible by . Otherwise there are two families, each of size greater than . Indeed .
We show that each family attains every label sum in the appropriate coset. Fix a label in family , and let be generated by all label differences there. For every the -sums of the labels fill . For each other occurring label reserve disjoint pairs with labels . Fewer than pairs suffice, and all required copies exist. Selecting one from each pair fills a coset of . If is the number of reserved pairs, then . Since the family has more than points and , there are enough unused points to add a fixed selection of points. All -term sums belong to , so the coset attained is precisely .
We distinguish two cases according to the order of the difference between the families. Let , and let be the order of in . If , set . Since , . Choose points in the first family and in the second. Their possible total labels include because . Then their horizontal sum is zero modulo and their height sum is .
For , we obtain a primitive row contradicting Lemma 3.4. There is a homomorphism taking to a unit. For each prime power exactly dividing , project onto a cyclic primary factor in which has that full prime-power order, then combine by the Chinese remainder theorem. The pullback to has the form . Choose and modulo each prime dividing but not . The form is primitive modulo . At a prime common to , its second coefficient is a unit since the character takes to a unit and reduces there to . At a prime dividing alone, cannot both vanish modulo that prime. At every remaining prime use the choice of . On family the form equals , constant because the character kills . Its fiber is a primitive row with more than points, contrary to Lemma 3.4. □
Lemma 3.6. Let be an integer and . If , a counterexample cannot have all but of its points in parallel primitive rows.
Proof. We first obtain an -point selection with the required height sum and a controlled horizontal sum. Let , , , and set
Since , our hypothesis gives , as required by Lemma 3.5. In every designated row of size delete points, less than in total. Make the projected selection of that lemma at . One designated row has selected points. Its original size satisfies . Fix the choices on every other row. If is the needed horizontal sum, the projected selection gives a -subset on this row with sum mod . We may now choose from all original points on this row.
We next preserve the horizontal sum modulo while retaining enough independent choices on this row. Represent its original horizontal coordinates in . Partition this interval into intervals of consecutive integers (the last may be shorter). Within each form disjoint packets of points until fewer than remain. At most points are left over, since . As , we can retain packets. There exists a -set on the row taking exactly half of every retained packet, since . Pair bijectively with . Their oriented differences sum to mod . Repeatedly discard nonempty zero-sums of differences modulo until fewer than remain (the partial-sum principle in ). Perform these remaining exchanges. This gives a -set of the required sum modulo and disturbs at most of the retained packets.
To attain the exact horizontal sum, fix everything except the choices of in each undisturbed packet. There are at least such packets. Indeed . Let be the stabilizer of the resulting horizontal sumset. If , each packet gives at least sums modulo . Indeed, order its points increasingly and move the selected indices from the first half to the last by individual unit steps, always keeping them distinct. The integer sums strictly increase over steps, and their diameter is less than . In the total sumset is aperiodic and proper. Kneser would give
which contradicts . Thus , so . The selection obtained from has sum congruent to modulo . Its difference from lies in the stabilizer. Consequently itself is attained. The fixed height sum then gives an -point zero-sum subset. □
Lemma 3.7. For , define
If , every primitive row of a counterexample contains fewer than points.
Proof. We show that such a row would concentrate almost all points on boundedly many parallel rows. Suppose a row contains at least points. Fix of size on that row and mark its row. While at least points are on unmarked rows take such points as . Their sizes add to at most , and ensures . Each set has size at least . By Lemma 3.2, some translate of contains more than members of . Mark its row. These are new points each time, so rows suffice to leave fewer than points. Apply Lemma 3.6. □
We now return to the box of Lemma 3.3. The bound on points in a primitive row will ensure that its long directions generate a subgroup with a sufficiently small quotient exponent. Retain the constants , , of that lemma and set
The integer is divisible by every possible stabilizer index up to . The numbers account for the points lost by fixing short coordinates of the box.
Lemma 3.8. Suppose . Some slice of the box in Lemma 3.3 has remaining coordinates and contains of the chosen points, where
Here is the volume of the slice.
Proof. We successively fix short coordinates while preserving the density and a lower bound for the number of points. After coordinates have been fixed, suppose the current box contains at least points and has density at least . If a side has length , fix the value of that coordinate which retains the most points. The volume is divided by , and the new slice contains at least a fraction of the points. Its density is therefore still at least , and its population exceeds . Repeat until all remaining sides have length at least . Fixing every coordinate would leave one site containing at least points, which is impossible. At the stopping step, , and every remaining side has length at least . □
We next bound the exponent of the quotient by the subgroup generated by the remaining directions. Translate the surviving box so that its base point is zero. We translate all of , since the sum of points is translation invariant. Let , , and set
If , Lemma 3.7 gives
Indeed, Smith normal form for the inverse image of in gives a basis in which with . The points lie on at most parallel primitive rows. One row contains at least points. If , this is at least , a contradiction. In particular , since for cyclic the invariant factors of have orders , forcing .
Lemma 3.9. Assume that the slice in Lemma 3.8 has remaining coordinates. Partition its points into disjoint sets whose sizes differ by at most one. In , consider the lines in the box obtained by varying only coordinate , and let . Keep the lines containing at least points of . On each retained line of population , prescribe an integer . If is the set of -coordinates on that line, the sum of the sets over all retained lines has stabilizer of index at most in .
Proof. We first estimate the total number of points on the retained lines. There are at most lines and , since . Discarding the shorter lines costs at most points. The retained lines thus contain points. Because there is another side of length at least , we also have
To bound the stabilizer index, let be the set of the distinct coordinate parameters on a retained line and consider its -sumset . Let be the stabilizer index of their total sumset. If , Kneser in gives
We exclude in the two cases and . Suppose first that . A line of population occupies at least residues. Its sums of points occupy at least residues. For , set aside one representative per residue and choose with and . Such exists since . Fix of the other points and vary of the representatives, using the restricted sum bound in . For the claim is immediate. The total growth is therefore at least
The middle inequality uses and . The last follows from (5), contradicting the upper bound .
Suppose next that , and let . On each line fix points, where . With , the remaining points give at least distinct -sums modulo . The increasing-index chain used in Lemma 3.6 has steps and diameter at most . Let and . Since ,
The right side is at least , since and . It is also at least , by (5) and . If , the sum of the 's is . Otherwise it is at least , since . Both alternatives contradict Kneser. Hence . □
Proof of Theorem 1.1. We first choose large enough to apply the preceding lemmas. The constants , and in (2) and (3) are independent of , with . Thus for all sufficiently large we have and .
We choose a total number of points divisible by . Suppose there were a counterexample for such an . Take the box of Lemma 3.8 and the disjoint point sets of the preceding lemma. On each retained line of population , choose the number of selected points to be a multiple of in . This is possible because . The largest and smallest allowed choices on that line differ by at least . Adding the corresponding arithmetic progressions over all lines gives every multiple of between the extreme totals. At least points lie on retained lines, so the interval of totals has width at least , lower endpoint at least , and upper endpoint at most . By (4), . The interval therefore contains a multiple of , and . Fix choices of giving that total.
We next show that the choices in attain all multiples of the stabilizer index in coordinate . Let be the stabilizer index given by Lemma 3.9. Since , it divides . On every retained line we can choose points whose -coordinate sum vanishes modulo . Indeed, and , so this follows by successive applications of the cyclic Erdős–Ginzburg–Ziv theorem. We remove disjoint zero-sum subsequences of length , with at least terms available before the last choice. Thus the sumset of the varying -coordinates meets . Its stabilizer is , so it contains all of . The other coordinates on each line are constant. Their contributions are multiples of because .
We combine these coordinate sums to cover a subgroup of . The are disjoint, so these choices can be made independently. In coordinate , the fixed contributions from the other sets belong to . Mapping the coordinate sums back into , we conclude that every element of
is the sum of exactly of the points.
It remains to complete the selection using points outside the chosen set of points. Since and , it follows that divides both and . Hence is a positive multiple of . There are at least points outside the chosen , and
Apply (1) to their images in . We obtain points whose sum belongs to . Choose points in the box with sum . The two choices are disjoint and contain points with sum zero. This proves the upper bound. Kemnitz’s lower bound gives equality. □
Remark. Theorem 1.1 holds for every , where one may take
Let . By (2), , , and . The recurrence gives
since and . In particular, and . Since , the definition in Lemma 3.7 gives . Thus
The elementary bound now gives . Since , every satisfies both size conditions in the proof.
The explicit threshold can be reduced by refining the estimates. We have chosen this value to keep the quantitative argument concise. We expect that further optimization, with the aid of large language models, could substantially reduce the threshold, while still leaving an extremely large bound. In our view, such improvements have limited theoretical significance unless can be reduced to a small integer.
Declaration of generative AI and AI-assisted technologies
The authors used a large language model during the preparation of this manuscript for assistance with mathematical exploration, checking arguments, organization and language editing. The authors are responsible for the mathematical statements, proofs, references and final text.
References
School of Mathematical Sciences, Tiangong University, Tianjin 300387, P.R. China Email address: [email protected]
Center for Combinatorics, Nankai University, Tianjin 300071, P.R. China Email address: [email protected]
References
- [1]N. Alon and M. Dubiner, Zero-sum sets of prescribed size, in Combinatorics, Paul Erdős is Eighty, Vol. 1, János Bolyai Math. Soc., Budapest, 1993, 33–50.
- [2]N. Alon and M. Dubiner, A lattice point problem and additive number theory, Combinatorica 15 (1995), 301–309.DOI
- [3]M. DeVos, L. Goddyn and B. Mohar, A generalization of Kneser’s addition theorem, Adv. Math. 220 (2009), 1531–1548.DOI
- [4]P. Erdős, A. Ginzburg and A. Ziv, Theorem in the additive number theory, Bull. Res. Council Israel Sect. F 10 (1961), 41–43.
- [5]W. Gao, On zero-sum subsequences of restricted size, J. Number Theory 61 (1996), 97–102.DOI
- [6]W. Gao, Note on a zero-sum problem, J. Combin. Theory Ser. A 95 (2001), 387–389.DOI
- [7]W. Gao and A. Geroldinger, Zero-sum problems and coverings by proper cosets, European J. Combin. 24 (2003), 531–549.DOI
- [8]W. Gao and A. Geroldinger, Zero-sum problems in finite abelian groups: a survey, Expo. Math. 24 (2006), 337–369.DOI
- [9]W. Gao, A. Geroldinger and W. A. Schmid, Inverse zero-sum problems, Acta Arith. 128 (2007), 245–279.DOI
- [10]W. D. Gao and R. Thangadurai, A variant of Kemnitz Conjecture, J. Combin. Theory Ser. A 107 (2004), 69–86.DOI
- [11]B. Girard, S. Griffiths and Y. O. Hamidoune, k-sums in abelian groups, Combin. Probab. Comput. 21 (2012), 582–596.DOI
- [12]P. Graham, The Harborth constant of Cₚ ⊕ Cₚ for every prime p ≥ 11, Zenodo preprint, 4 November 2026, doi:10.5281/zenodo.23968472, arXiv:2611.00487.
- [13]B. Green and I. Z. Ruzsa, Freiman’s theorem in an arbitrary abelian group, J. Lond. Math. Soc. 75 (2007), 163–175.DOI
- [14]P. Guillot, L. E. Marchan, O. Ordaz, W. A. Schmid and H. Zerdoum, On the Harborth constant of C₃ ⊕ C₃ₚ, J. Théor. Nombres Bordeaux 31 (2019), 613–633.
- [15]H. Harborth, Ein Extremalproblem für Gitterpunkte, J. Reine Angew. Math. 262/263 (1973), 356–360.DOI
- [16]A. Kemnitz, Extremalprobleme für Gitterpunkte, doctoral dissertation, Technische Universität Braunschweig, 1982.
- [17]A. Kemnitz, On a lattice point problem, Ars Combin. 16-B (1983), 151–160.
- [18]L. E. Marchan, O. Ordaz, D. Ramos and W. A. Schmid, Some exact values of the Harborth constant and its plus-minus weighted analogue, Arch. Math. 101 (2013), 501–512.arxiv.org/abs/1308.3315
- [19]G. Petridis, New proofs of Plünnecke-type estimates for product sets in groups, Combinatorica 32 (2012), 721–733.
- [20]C. Reiher, On Kemnitz’ conjecture concerning lattice-points in the plane, Ramanujan J. 13 (2007), 333–337.DOI
- [21]C. Reiher and T. Schoen, Note on the theorem of Balog, Szemerédi, and Gowers, Combinatorica 44 (2024), 691–698.DOI
- [22]L. Rónyai, On a conjecture of Kemnitz, Combinatorica 20 (2000), 569–573.DOI
- [23]S. Savchev and F. Chen, Long zero-free sequences in finite cyclic groups, Discrete Math. 307 (2007), 2671–2679.DOI
- [24]S. Savchev and F. Chen, Long n-zero-free sequences in finite cyclic groups, Discrete Math. 308 (2008), 1–8.
- [25]B. Sury and R. Thangadurai, Gao’s conjecture on zero-sum sequences, Proc. Indian Acad. Sci. Math. Sci. 112 (2002), 399–414.
- [26]R. Thangadurai, On a conjecture of Kemnitz, C. R. Math. Rep. Acad. Sci. Canada 23 (2001), 39–45.