## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Ratings:

383 pages6 hours

This book develops the thesis that structure and function in a variety of condensed systems - from the atomic assemblies in inorganic frameworks and organic molecules, through molecular self-assemblies to proteins - can be unified when curvature and surface geometry are taken together with molecular shape and forces. An astonishing variety of synthetic and biological assemblies can be accurately modelled and understood in terms of hyperbolic surfaces, whose richness and beauty are only now being revealed by applied mathematicians, physicists, chemists and crystallographers. These surfaces, often close to periodic minimal surfaces, weave and twist through space, carving out interconnected labyrinths whose range of topologies and symmetries challenge the imaginative powers.

The book offers an overview of these structures and structural transformations, convincingly demonstrating their ubiquity in covalent frameworks from zeolites used for cracking oil and pollution control to enzymes and structural proteins, thermotropic and lyotropic bicontinuous mesophases formed by surfactants, detergents and lipids, synthetic block copolymer and protein networks, as well as biological cell assemblies, from muscles to membranes in prokaryotic and eukaryotic cells. The relation between structure and function is analysed in terms of the previously neglected hidden variables of curvature and topology. Thus, the catalytic activity of zeolites and enzymes, the superior material properties of interpenetrating networks in microstructured polymer composites, the transport requirements in cells, the transmission of nerve signals and the folding of DNA can be more easily understood in the light of this.

The text is liberally sprinkled with figures and colour plates, making it accessible to both the beginning graduate student and researchers in condensed matter physics and chemistry, mineralogists, crystallographers and biologists.

Publisher: Elsevier ScienceReleased: Nov 19, 1996ISBN: 9780080542546Format: book

*1996 *

During the latest decade we have worked on periodic surfaces with zero average curvature and their significance in chemical structures, ranging from atomic and molecular arrangements in crystals to complex self-assembled colloidal aggregates. This approach has proved to be fruitful, not only in the determination of complex structures, but also in the understanding of phase behaviour and relations between structure and physical properties. Our aim has been to summarise our own understanding of this growing field, and to provide a complete description of relevant shapes and the forces behind their formation.

This book deals with the role of curvature, a neglected dimension, in guiding chemical, biochemical and cellular processes. The curved surfaces that concern us might be those traced out by the head groups of phospholipid molecules that spontaneously self-assemble to form membranes and other building blocks of biology. Or they can be the surfaces of proteins involved in catalysis. They are provided in abundance *par excellence *by inorganic chemistry. In biology these dynamic entities have a marvellous capacity for self-organisation and self-assembly which is beginning to be understood. They transform from one shape to another under the influence of the forces of nature with an astonishing ease that allows them to manage resources, direct complex sequences of reactions, and arrange for delivery, all on time. Shape determines function, and the energetics of function dictates the optimal structure required. At least that is our thesis.

The cognition and recognition of shape and form are one of the earliest tasks presented to the brain. Shape and form are so much a part of our mental processes that we tend to take them for granted. Almost any word in any language that describes objects conjures up an image that involves form. And indeed one of the deepest expressions of our sense of being is representational art. Painting and sculpture deal exclusively with colour, shape and form. Yet despite the vaunted successes of physics and mathematics that underlies modern science, science remains antithetic to art because it reduces diversity to too sterile order through the imposition of Euclidean symmetry.

According to conventional texts, forces act between point atoms, spheres, cylinders and planes in a kind of pythagorean and ptolemaic imperative that ignores curvature. There is nowhere an awareness that shape may have a role to play, except to please the eye. Nature ever geometrised, said somebody. True. But it has good reasons.

In cell and molecular biology where mechanisms of enzyme action are not understood and attributed to some kind of Maxwell demon, all is specificity, and the lipids of membranes serve to do no more than act as a passive matrix for proteins and as a protection for the procreation of a uni-dimensional, machine like and stolidly boring DNA. There is more to it than that.

The thesis of this book is that two circumstances may have contributed to our present situation. The one has to do with the forces acting between chemical assemblies, and the interplay between these forces, set by the environment in which they work, and curvature. The other has to do with the absence of any language describing shapes of physically associated assemblies that are part of the subject of cellular and molecular biology. When shape is taken into account one comes to the realisation that curvature, and forces, set by constraints, are meaningful thermodynamic variables, (derived from classical thermodynamics). The key problem in the reductionist chain is how to build a statistical mechanism that uses a language of shapes. This language draws on topology and differential geometry. What we will attempt to show is that once that language is learnt, the world begins to take on a richer and more colourful unity. Through a consideration of minimal surfaces and other shapes the bewildering chaos of nature makes more sense.

We are convinced that a structural description based on curvature is useful in physical and biological sciences, and the numerous examples presented here support that view. Finally, we hope that our speculations on the role of these shapes in chemical reactions and in molecular organisation in living systems will inspire new work in this field.

**Chapter 1 **

**The Mathematics of Curvature **

This book deals with shape and form, and especially the role of curvature in the natural sciences. Our search is for a connection between structure and function posed by D’Arcy Thompson in his famous book On Growth and Form

[**1] almost a century ago. Our theme will be that curvature, a neglected dimension, is central. Some of the curved surfaces that will preoccupy us and recur are shown in the Appendix to this Chapter. The reader is invited to pursue them at once. They are not just computer generated art or mathematical abstractions, and will be seen later to be ubiquitous in nature. They represent situations as diverse as: **

• equipotential surfaces dividing space between the atoms of a crystal

• real structures formed spontaneously by the constituent molecules of biological membranes

• the shapes of bio-macromolecules, from proteins to starch

Euclidean geometry underlies practically all of science and our intuition has depended on it. The shapes provided: planes, cylinders, spheres, polyhedra, all have constant or even zero curvature. Only in theoretical physics, in subjects like general relativity where the curvature of space-time is essential, has non-Euclidean geometry and especially so-called hyperbolic geometry played any part in the scheme of things. The scientific community has been prepared to leave such matters to physicists alone. It can do so no longer, and the idea of curvature is becoming an essential tool to the understanding of many phenomena.

This Chapter is concerned with some of the mathematical tools required to describe special properties of curved surfaces. The tools are to be found in differential geometry, analytical function theory, and topology. General references can be found at the end of the Chapter. The reader uninterested in the mathematics can skip the equations and their development. The ideas we want to focus on will be clear enough in the text. A particular class of saddle-shaped (hyperbolic) surfaces called *minimal surfaces *will be treated with special attention since they are relatively straightforward to treat mathematically and do form good approximate representations of actual physical and chemical structures.

The concept of curvature was developed by Isaac Newton in the middle of the 17th century, as a natural extension to his work on the calculus. At that time, the determination of the perimeter of planar curves and the area under curves were major problems. In particular, Newton’s new analytical tools allowed him to determine the quadrature

(area) of a circle. It occurred to Newton that the radius of the circle of best fit to an arbitrary planar curve at all points on the curve was a useful measure, for which he coined the term crookedness

[**2]. This is curvature (Fig. 1.1). **

**Figure 1.1 **The curvature of a planar curve at a point (P) is equal to the reciprocal of the radius of the circle of best fit to the curve at P, r.

The curvature of a planar curve relates arc length along the curve to changes of tangent vector (**Fig. 1.2). **

**Figure 1.2 **Tangents TP and QT at two points, P and Q, on a planar curve.

The tangents TP and QT in **Fig. 1.2 subtend angles ψ, ψ+δψ with the x-axis, so that δψ is the average curvature of the planar curve along the arc PQ. The curvature at the point P is defined to be the limit of this expression as Q approaches P, i.e**

If PQ is the arc of a circle of radius *r*, the angle δ*ψ *between the tangents at P and Q is equal to the angle subtended at the centre of the circle by the arc PQ, so that δ*s=r*δ*ψ*. The curvature is constant at all points of a circle, and the radius is equal to the reciprocal of the curvature (**Fig. 1.1). If the curve is described in cartesian coordinates by a function y=y(x): **

so that:

The curvature, *κ*, is thus given by the expression:

**(1.1) **

; *i.e*is zero and therefore the curvature is zero in these cases. The sign of the curvature signifies the convex or concave nature of the curve. It is also related to the side of the curve at which the centre of the circle of best fit is located (*cf*. **Fig. 1.3). **

**Figure 1.3 **The radii of curvature, rP and rQ at two points, P and Q, on a planar curve. The centres of the circles of best fit to the curve at P and Q lie on opposite sides of the curve and the curvature changes sign at the point of inflection on the curve between these points. The curvature at Q is positive and at P it is negative.

The *curvatures *of a *surface *are more complex entities, but can be understood as a generalisation of the curvature of planar curves. Imagine a plane containing a point P on the (smooth) surface, which contains the vector (n) passing through P, normal to the surface (**Fig. 1.4). **

**Figure 1.4 **The intersection of a surface with the plane containing the normal vector (n) to the surface at the point P.

The intersection of this plane with the surface is clearly a planar curve, whose curvature at P can be evaluated as described above. This curvature is equal to the value of the *normal curvature, κn*, at P in the direction prescribed by the orientation of the plane. Now let the plane rotate about an axis coincident with the normal vector, n. The normal curvature will vary periodically, so it must attain maximum and minimum values. These values are defined to be the *principal curvatures, κ*1 and *κ*2, of the surface at P (**Fig. 1.5). The directions at which these extremes occur are referred to as the principal directions at P. In special cases, all these curves of intersection are of equal curvature (e.g. a point on a sphere), the point is an umbilic, and principal directions cannot be defined. If the normal curvatures at the umbilic are zero (so that all the intersection curves are straight lines) the surface is locally planar at that point, which is then called a flat point. At regular points (excluding umbilics) the principal directions are orthogonal. **

**Figure 1.5 **The extrema of normal curvatures define the principal curvatures of a surface.

The principal curvatures can be combined to give two useful measures of the curvature of the surface, the *Gaussian curvature *(*K*) and the *mean curvature *(*H*):

**(1.2) **

The surfaces in **Fig. 1.5 have (a) positive Gaussian curvature, (b) zero Gaussian curvature and (c) negative Gaussian curvature. **

The Gaussian curvature has the dimensions of inverse area and the mean curvature has dimensions of inverse length. The *topology *of the surface (introduced below) is related to a (dimensionless) measure of the integral geometry of the surface, the *integral curvature*.

The Gaussian curvature and integral curvature bear a fascinating relation to the normal vectors on the surface, and belong to the realm of *intrinsic geometry, i.e*. the geometry that can be deduced without reference to the space within which the surface is embedded. Some further results on the intrinsic geometry of surfaces will be needed throughout the book. We outline them briefly below.

The Gaussian curvature has a number of interesting geometrical interpretations. One of the more striking is connected with *the Gauss map *of a surface, which maps the surface onto the unit sphere. The image of a point P on a surface *x *under the mapping is a point on the unit sphere. This point is given by the intersection of the unit normal n to the surface at P with a unit sphere centred at P. The Gauss map of the surface *x *is the collection of all such points on the sphere, generated by sliding the surface through the centre of the (fixed) sphere (**Fig. 1.6). If a closed curve on a surface is traversed in the opposite sense on the sphere under the mapping, the surface is saddle-shaped, and the Gaussian curvature is negative. **

**Figure 1.6 **The Gauss map of a surface. The normal vectors in the triangular ABC region of the saddle-shaped surface define a region on the unit sphere, A’B’C, given by the intersection of the unit sphere with the collection of normal vectors (each placed at the centre of a unit sphere) within the ABC region. Notice that for the example illustrated the bounding curve on the surface and on the unit sphere are traversed in opposite senses. This is a necessary feature of saddle-shaped surfaces, with negative Gaussian curvature.

Clearly the spherical image under the Gauss map of a highly curved surface patch will be larger than that of less curved patches of the same area, since the divergence in direction spanned by the normal vectors is wider for the highly curved patch. An extreme example is the plane, which is mapped onto a single point, whose location depends on the orientation of the plane.

. An alternative definition of the Gaussian curvature follows from this result. Imagine shrinking the region progressively to an infinitesimal area about a point. In the limit the quotient of the area of the surface element and its spherical image is 1*/K*. If the Gauss map of a surface comprises only a single point (*e.g*. the plane) or a curve (e.g. the cylinder), the Gaussian curvature is zero at all points on the surface.

The Gaussian curvature, *K*, is a bending invariant. This means that if we can bend a simply connected surface *x *into another simply connected surface *y *without stretching or tearing, there exists a continuous transformation from *x *to *y *that preserves the Gaussian curvature at every point. For example, the plane, with Gaussian curvature, *K ≡ *0, is easily rolled into a cylinder for which also *K ≡ *0. On the other hand there is no way to form a sphere (*K *constant, but strictly positive) from either cylinder or plane without stretching, tearing or gluing. Surfaces that are related by a curvature-preserving transformation (like the plane and the cylinder) are called *isometric*.

Another entity that we shall need belongs to the realm of intrinsic geometry: *geodesic curvature*. Consider a surface *x*, a point P on *x *and a curve *ξ *on *x *passing through P. The curvature vector of *ξ *at P joins P to the centre of curvature of *ξ*. This curvature vector may be decomposed into mutually orthogonal components. These components are given by projection of the

curve *ξ *onto two orthogonal planes: (i) the tangent plane to *x *at P and, (ii) the plane containing the normal vector to *x *at P and the tangent vector of *ξ *at P. The curvature of the latter projection is the normal curvature, *κn*, introduced in **section 1.3. The geodesic curvature, κg of ξ at P on x is equal to the curvature of the projection of ξ onto the tangent plane to x at P (Fig. 1.7). If the geodesic curvature is zero, the curvature of ξ is identical to the normal curvature. A curve whose geodesic curvature is zero everywhere is called a geodesic, and it is (locally) the shortest distance between two points on the surface. Along geodesic curves, the normal vectors to the geodesic coincide with the normal vectors to the surfaces. An infinite number of geodesics passes through any point, one for every direction emanating from the point. Geodesics on curved surfaces are rarely straight lines. Geodesics on a curved surface linking two points can be constructed by stretching a string (constrained to lie on the surface) between the points - the path taken up by the string will always follow a geodesic. **

**Figure 1.7 **Decomposition of a curve in a surface (left) into orthogonal geodesic and normal curvatures (right).

One further measure of the bending of curves needs mention. A nonplanar space curve exhibits curvature (which is measured by the radius of the circle of best fit to the curve) and *torsion*.

The torsion of a curve describes its pitch: a helix exhibits both constant curvature and torsion. Its curvature is measured by its projection in the tangent plane to the curve - which is a circle for a helix - while its torsion describes the degree of non-planarity of the curve. Thus a curve on a surface (even a geodesic), generally displays both curvature and torsion.

The measure of torsion is unambiguous for an isolated curve pictured in three-dimensional space. However, the torsion of a curve lying in a surface has a more complicated property related to the geometry of the surface. The *geodesic torsion *(*τg*) is a further measure of the local bending of a surface curve complementary to the normal and geodesic curvatures. The geodesic torsion at a point on a surface in a certain direction is equal to the torsion of the geodesic on the surface through that point in that direction. This can be stated more formally as follows. A triple of orthogonal vectors can be defined at any point on a curve lying in a surface. This triple contains the normal vector to the surface at that point, n, the tangent vector to the curve at that point, t, and a vector orthogonal to both of those vectors, known as the *geodesic normal vector*, u = n × t. The rate of change of n with arc length along the curve projected onto t is equal to the normal curvature. The rate of change of n projected onto u is equal to the geodesic torsion, *tg*. The geodesic torsion thus complements both the geodesic and normal curvatures in a natural way, although (along with the normal curvature) the geodesic torsion is not a concept of intrinsic geometry. All three measures of bending of curves on surfaces can be unified by the Bonnet-Kovalevsky formulae:

Both the normal curvature and the geodesic torsion of a curve on a surface depend on the variation of the normal vectors to the surface along the curve, which implies some unexpected results. For example, a straight line certainly displays neither curvature nor torsion. However, the geodesic torsion of a straight line vanishes only if the line is embedded in a surface of zero Gaussian curvature. The normal curvature and geodesic torsion of a curve on a surface are related to the principal curvatures of the surface (*κ*1 and *κ*2) and the angle the curve subtends with a principal direction, *φ*:

**(1.3) **

**(1.4) **

For a surface characterised by *κ*1 = *κ*2, the Gaussian curvature is simply related to the normal curvature and geodesic torsion:

**(1.5) **

In this case, the magnitude of the geodesic torsion at a point on a straight line lying in the surface is equal to the magnitude of the principal curvatures of the surface at that point.

The Gauss-Bonnet theorem is a profound theorem of differential geometry, linking global and local geometry. Consider a surface patch *R*, bounded by a set of *m *curves *ξi*. If the edges *ξi *meet at exterior angles *θi *and they have geodesic curvature *κ*g(*si*) where *si *labels a point on *ξi *then the theorem says

**Figure 1.8 illustrates the case for a surface patch consisting of four boundary arcs, ξ1=AB, ξ2 =BC, ξ3=CD and ξ4 = DA. **

**Figure 1.8 **Four arcs belonging to a surface. From the Gauss-Bonnet theorem, the integral curvature within the region of the surface bounded by the arcs (ABCD) is determined by the vertex angles (*θi*) and the geodesic curvature along the arcs AB, BC, CD and DA.

Choose a triangle traced on a surface, whose three edges are geodesics. From the theorem, we have

**(1.7) **

The external angles, *θi*, are related to the internal angles, *αi*, by *θi *= π − *αi *so that the area of the geodesic triangle is:

**(1.8) **

Thus, for a Euclidean triangle (which is located on a surface of zero Gaussian curvature, such as the plane), the sum of the vertex angles of a triangle is indeed π. However, if the triangle decorates a surface of negative Gaussian curvature, the sum of the angles is less than π, if the integral surface has positive integral curvature, the sum of the angles exceeds π. The angle excess

(*i.e*. its difference from π) is thus a measure of the integral curvature within the region bounded by the geodesic edges.

Geometry, differential or otherwise, deals with the metric relationships of rigid objects. There are some fundamental aspects of shapes that are preserved if the objects studied consist of stretchable rubber sheets. A rubber sphere may be deformed into an ellipsoid, or a long, narrow cylinder with caps, or indeed any globular object (**Fig 1.9). **

**Figure 1.9 **A sphere can be stretched and bent (without any rupture or fusion of the surface) into an infinite variety of globular surfaces, all topologically equivalent.

Similarly, a torus can be deformed into any two-sided surface containing a single handle, or a single hole (**Fig 1.10). **

**Figure 1.10 **A donut-shaped torus (with a single hole) can be deformed into any two-sided surface containing a single handle, such as a cup.

Surface topology is rubber sheet

geometry, since only those geometrical characteristics of the surface that are maintained upon stretching or squeezing are relevant. The usual geometrical notions of area, length, etc., are excluded from topological analysis.

Suppose a surface, x, is facetted, subdividing x into a number of faces, edges (bounding the faces) and vertices. Denote the number of faces by F, the number of their edges by E and the number of vertices by *V*. Descartes, and later Euler, discovered that (*F-E+V*) =2 for all polyhedra (**Table 1.1). **

**Table 1.1 **

**Relation between numbers of faces ( F), edges (E) and vertices (V) of conventional polyhedra **

This property holds because (*F-E+V*) is a topological characteristic, dependent only on the topology of the facetted surface. Since all polyhedra are topologically equivalent to the sphere (**Fig 1.9), ( F-E+V) is conserved. The value of this integer is known as the Euler-Poincaré characteristic χ(x): **

**(1.9) **

Now imagine a net lying in a smooth surface. This net also facets the surface into curved faces, curved edges and vertices. If each node on the net has *z *edges (so that the net is z-connected), and the ring size of each ring in the net is *n*, Euler’s relation for the net is:

This result implies that the ring size and connectivity of a network determine the topology of the surface which contains that network. This allows for simple characterisation of cage, sheet and framework nets, distinguishable by the value of their Euler-Poincaré characteristic (**Table 1.2). **

**Table 1.2 **

**Relation between average ring size, n, connectivity, z, and Euler-Poincaré characteristic per vertex, χ/V, for a range of networks that are regular tessellations of surfaces. The nature of the network is set by the value of χ/V: cages, planar networks and three-dimensional frameworks are characterised by positive, zero and negative χ/V respectively **

Another topological characteristic, the genus, *g*(*x*), of a surface, is a measure of its connectedness. It is equal to the number of holes or handles in the surface and simply related to the Euler-Poincaré characteristic by

**(1.11) **

(The equation applies only to ‘orientable’ surfaces, those with distinct sides. This excludes one-sided surfaces, such as the Möbius strip.) Thus, a sphere has genus zero. A torus (or a sphere with one handle) has genus one, and so on.

Remarkably, the topology is linked to the integral curvature of a surface by the simple equation:

**(1.12) **

This means that all surfaces with the same number of handles or holes have the same integral curvature! In other words, no amount of bending or squeezing of a surface can add any net integral Gaussian curvature to that displayed by the simplest topologically equivalent surface. For example, although regions of negative Gaussian curvature can be formed in a sphere by squeezing the surface to produce a region which is saddle-shaped, this contribution to the integral curvature will always be compensated by a corresponding positive integral curvature in other regions.

A more general, global version of the Gauss-Bonnet theorem can now be stated: Let *x *be an oriented surface and *R *be a bounded region of *x*. As before, let the boundary of *R *be the union of *m *simple curves *ξi *that do not self-intersect, and let *θi *be the external angles at the *m *vertices.

Then we have

It is obviously impossible to offer a complete catalogue of curves. Similarly, no comprehensive list of surface forms can be drawn up. The language of surface shape is a rich one: some familiar forms like the sphere and the plane, are deeply imbued in our consciousness, while others remain difficult to describe and visualise in terms that are intuitively reasonable to all of us raised on the limited vocabulary afforded by the simpler forms. But some attempt at exhaustive classification is necessary: the wealth of form in natural structures draws on the richness of abstract form, so that if we are to understand natural structures, it is necessary to obtain as full an intuition about surface forms as possible.

To do this we must draw on non-Euclidean geometries. These different geometries emerge if the parallel postulate

of Euclid’s Elements

is no longer taken as axiomatic. Following the work of Bolyai, Lobachevsky, Gauss and Riemann early last century it became clear that if the parallel postulate is relaxed, three quite distinct geometric classes exist. The three classes are elliptic (single and double), parabolic (Euclidean) and hyperbolic. Riemann’s great contribution to geometry lay with his program of abstracting and relating the notion of form to the concept of differential form. The Riemannian approach classifies a geometry by its functional (pointwise) structure - characterised by the curvatures and metric. The triad of local surface geometries is characterised by Gaussian curvature: elliptic shapes have positive Gaussian curvature, Euclidean shapes have zero Gaussian curvature and hyperbolic shapes have negative Gaussian curvature.

The vocabulary of surface forms can be developed by marrying this local geometry to (global) topology. Any surface form can be characterised by its intrinsic shape (its metric and curvature) and its global (extrinsic) embedding in space. Since natural structures are embedded in our particular space, we are concerned in this book with two-dimensional (surface) embeddings in three-dimensional Euclidean space. (We leave the cosmologists to debate the nature of the deviations from Euclidean space due to matter. However, it is certain that at the length scales we deal with here, the space we live in can be safely approximated by Euclidean space.) Thus, although the embedding space is always approximately Euclidean, its intrinsic geometry may be locally Euclidean (parabolic) or non-Euclidean, *viz*. elliptic or hyperbolic.

In the vicinity of elliptic points, the surface can be fitted to an ellipsoid, whose radii of curvature are equal to those at that point (**Fig. 1.5a). The surface lies entirely to one side of its tangent plane, it is synclastic and both curvatures have the same sign. About a parabolic point, the surface resembles a cylinder, of radius equal to the inverse of the single nonzero principal curvature (Fig. 1.5b). Hyperbolic (anticlastic) points can be fitted to a saddle, which is concave in some directions, flat in others, and convex in others (Fig. 1.5c). At hyperbolic points, the surface lies both above and below its tangent plane. **

The most familiar surfaces have constant Gaussian curvature over the surface. In our space this is only possible in elliptic and parabolic cases: the sphere and the cylinder and plane respectively. It is impossible to form hyperbolic surfaces of constant (negative) Gaussian curvature without singularities. In general surfaces contain elliptic, parabolic and hyperbolic regions. The average geometry

of a surface can be characterised by the average value of its Gaussian curvature, <*K*>. This is equal to the integral curvature divided by the surface area *A*:

**(1.14) **

The integral curvature of a surface is linked to the Euler-Poincaré characteristic of that surface (*χ*) by **eq. (1.12). This allows the average geometry of orientable surfaces to be related to the number of holes or handles, characterised by the surface genus, g, and the area of the surface, A, by the relation: **

**(1.15) **

Thus, surfaces that are free of holes or handles like the sphere have positive <*K*> and *χ; *they are elliptic. Surfaces with a single handle or a hole (a donut-shaped torus, or - plus or minus a single point - a cylinder or a plane) are on average parabolic, or two-dimensionally Euclidean (*χ*=0). All surfaces with more than one hole (or handle), are hyperbolic (negative <*K*>,

You've reached the end of this preview. Sign up to read more!

Page 1 of 1

Close Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Loading