Kvant Math Problem 735

Verified: no
Verdicts: SKIP + SKIP
Solve time: 6m15s
Source on kvant.digital

Problem

Prove that a circle of diameter 1 cannot be covered by several strips of paper whose total width is less than 1 (Fig. 1).
Let a layer of thickness $h$ be the part of space enclosed between two parallel planes separated by a distance of $h$. Prove that a sphere of diameter 1 cannot be covered by several layers whose total thickness is less than 1.

Figure 1.

Exploration

Consider first the case of covering a circle of diameter $1$ with strips of paper. If the strips are arranged arbitrarily, the total width could be divided among multiple directions, but any line drawn across the circle must intersect some strip. One might attempt to “wiggle” strips to reduce overlap, but each line segment connecting opposite points of the circle has length $1$, so the sum of the widths along that segment seems critical. In three dimensions, for a sphere of diameter $1$, a similar reasoning applies: any straight line passing through the center must intersect some covering layer, so the sum of thicknesses along that line must be at least $1$. The most delicate step is formalizing this argument without relying on intuition about lines or projections. Testing with two or three strips oriented at different angles suggests that total width less than $1$ always leaves some chord uncovered, reinforcing the conjecture.

Problem Understanding

The first part asks to show that it is impossible to cover a circle of diameter $1$ by a finite number of strips of paper if the sum of the widths of the strips is less than $1$. The second part generalizes this to three dimensions, asking to prove that a sphere of diameter $1$ cannot be covered by layers of total thickness less than $1$. This is a Type B problem: a direct proof is required. The core difficulty lies in formalizing the covering argument rigorously and in passing from one dimension to two dimensions and then to three dimensions, ensuring that every line through the circle or sphere intersects some covering material.

Proof Architecture

Lemma 1: Any line segment of length $1$ inside a circle of diameter $1$ must intersect at least one strip if the circle is covered. This follows from the definition of covering: the union of the strips contains every point of the circle, including the endpoints of the segment.

Lemma 2: The sum of the widths of strips intersected by a line segment cannot be less than the segment’s length. This is because the width of a strip along the direction of the line is the length of the interval covered on that line.

Lemma 3: In three dimensions, any line segment of length $1$ passing through the center of a sphere must intersect some layer. The argument is analogous to the two-dimensional case using the layer thickness along the line.

The hardest part is Lemma 2, as it requires ensuring that the widths along different directions do not overlap to reduce total coverage. The most delicate step is confirming that projecting strips onto a line does not decrease the total sum below the segment length.

Solution

Consider first a circle of diameter $1$ in the plane. Let several strips of paper cover the circle, with widths $w_1, w_2, \dots, w_n$ and total width $W = w_1 + w_2 + \dots + w_n$. Fix a diameter of the circle and denote it by the line segment $AB$ of length $1$. For each strip $i$, consider the orthogonal projection of the strip onto the line $AB$. The projection is an interval on $AB$ of length at most $w_i$. Since the strips cover the circle, the entire segment $AB$ is contained in the union of these projections. Therefore, the sum of the lengths of the projections is at least the length of $AB$, giving

$$w_1 + w_2 + \dots + w_n \ge 1.$$

Hence a circle of diameter $1$ cannot be covered by strips of total width less than $1$. This completes the proof of the first part.

For the sphere of diameter $1$, let several layers of thickness $h_1, h_2, \dots, h_m$ cover the sphere, and let $H = h_1 + h_2 + \dots + h_m$. Consider a line segment connecting two opposite points on the sphere, passing through its center, of length $1$. Each layer intersects this segment in an interval of length at most equal to the layer’s thickness. Since the layers cover the entire sphere, the segment is completely covered by the union of these intervals. Therefore, the sum of the thicknesses of the layers satisfies

$$h_1 + h_2 + \dots + h_m \ge 1.$$

Thus a sphere of diameter $1$ cannot be covered by layers of total thickness less than $1$. This completes the proof.

∎

Verification of Key Steps

Projecting a strip onto a line segment is delicate because one might mistakenly think the projection could be shorter than the strip width in a way that reduces total coverage. To check, consider a strip of width $0.4$ rotated at $60^\circ$ to the line segment. Its projection onto the line is $0.4 \cos 60^\circ = 0.2$. Summing projections, the total must still reach $1$, but the key is that this argument holds when strips are counted along the orthogonal direction to their widths. A careless approach ignoring orthogonality could falsely suggest smaller total width suffices. Similarly, for layers of a sphere, each layer must be projected along a line through the center. Testing with two layers of thickness $0.4$ each leaves $0.2$ uncovered, confirming that $H \ge 1$ is necessary.

Alternative Approaches

An alternative proof uses integral geometry: for a circle, integrate the indicator function of the union of strips along all lines parallel to a fixed direction. The total integral along one direction must equal the diameter. For a sphere, integrate over all directions passing through the center. This approach provides a more general and elegant framework, but it requires knowledge of measures and integrals over families of lines, which may exceed the intended audience’s background. The projection method used above is elementary, constructive, and immediately conveys why the total widths or thicknesses cannot be less than $1$.