experiments, such as the European Space
Agency’s Eddington and NASA’s Kepler
missions, will also search for extrasolar plan-
ets through their transit signatures. Avoiding
the data deterioration caused by the Earth’s
atmosphere, these aim to locate planets as
small as, or smaller than, the Earth. The
success of Konacki et al.
even greater enthusiasm for the promising
projects soon to come.
Timothy M. Brown is at the High Altitude
Observatory, National Center for
Atmospheric Research, 3450 Mitchell Lane,
Boulder, Colorado 80303, USA.
1. Udalski, A. et al. Acta Astron. 52, 1–37 (2002).
2. Udalski, A. et al. Acta Astron. 52, 115–128 (2002).
3. Konacki, M., Torres, G., Jha, S. & Sasselov, D. D. Nature 421,
4. Mayor, M. & Queloz, D. Nature 378, 355–359 (1995).
5. Goldreich, P. & Tremaine, S. Astrophys. J. 241, 425–441
6. Murray, N., Hansen, B., Holman, M. & Tremaine, S. Science
279, 69–72 (1998).
7. Henry, G., Marcy, G., Butler, R. & Vogt, S. Astrophys. J. 529,
8. Charbonneau, D., Brown, T., Latham, D. & Mayor, M.
Astrophys. J. 529, L45–L48 (2000).
9. Lissauer, J. Nature 419, 355–358 (2002).
ouseholds in many countries have
become smaller in recent decades.
Between 1970 and 2000, the average
number of occupants in households in less
developed countries fell from 5.1 to 4.4. And
in more developed nations, the decrease was
from 3.2 to 2.5 people per household over the
same period (the decline began earlier; Fig.
1). From their analysis of household dynam-
ics in biodiversity ‘hotspot’ areas, Liu and
now argue (page 530 of this issue)
that the decline in household sizes has un-
intended negative effects. The global human
population has risen, not fallen, so smaller
households means more households — and
a higher demand for natural resources. This
is in addition to the increased demand result-
ing purely from population growth.
Even before the writings of Thomas
Malthus in the late eighteenth century, the
balance between population and natural
resources was a recurrent theme. Since
ancient times, statesmen and philosophers
have expressed opinions about such issues as
the optimum number of people and the dis-
advantages of excessive population growth
Although some theorists see population
expansion in a positive light
, there is
increasing concern about the negative conse-
quences for resources
. Other things being
equal, a larger population implies a greater
demand for food, water, arable land, energy,
building materials, transport and so on — a
link that was first quantified some 30 years
. A population’s age structure also influ-
ences economic growth and hence resource
use: a rapid growth of the young age seg-
ments decelerates economic growth
More recently, scholars have acknowl-
edged that another demographic variable —
the number of households — also has an
important role in resource consumption
Even when the size of a population remains
constant, more households imply a larger
demand for resources. Household members
share space, home furnishings, transporta-
tion and energy, leading to significant
economies of scale. For instance, two-person
households in the United States in 1993–94
used 17% less energy per person than one-
To appreciate the different effects of pop-
ulation size and number of households on
resource consumption on a larger regional
scale, consider the following example
more developed regions, energy consump-
tion increased by 2.1% per year over the
period 1970–90. Population growth can
explain 0.7 percentage points of this growth
in energy usage, while changes in per capita
energy use explain the remaining 1.4 points.
However, an alternative analysis decompos-
es the growth in energy consumption into a
factor that describes the growth in number
of households and a factor describing per
household energy use. This analysis shows
that the household growth factor explains
1.6 percentage points of the energy-
consumption increase — more than twice
as much as the population growth factor.
Liu and colleagues
now draw our atten-
tion to household dynamics in biodiversity
hotspot areas — regions that are rich in
endemic species and threatened by human
activities. They find that, during the years
1985–2000, the number of households in 76
hotspot countries increased by 3.1% per year,
substantially faster than did the population
(1.8% per year). So, average household size
fell by about 1.3% per year. These changes
relate to the group of 76 countries as a whole.
For individual hotspot countries, more than
80% showed a pattern of greater growth in
household numbers than in population. In
65 non-hotspot countries, however, popula-
tion increased at roughly the same tempo as
household numbers during 1985–2000.
Many of the world’s most populated coun-
tries are hotspot countries (such as China,
India, Indonesia, Brazil and Bangladesh).
And most of the hotspot countries studied by
Liu et al. (65 out of 76) belong to the group of
less developed nations. We know that falling
birth rates were an important driving force
behind reductions in average household size
in less developed countries in the 1990s (ref.
12). Despite these falling birth rates, however,
the population in such countries did increase
(because of decreased death rates, for
instance). All of this might explain why
increases in the number of households were
relatively pronounced in hotspot countries
Liu et al. also refer to projections of popu-
lation size and the number of households
over the next 15 years. These projections
suggest that the divergence in population
growth and household numbers will become
more pronounced. So, the authors argue, it
is crucial to consider average household size
when assessing threats to biodiversity.
Quantifying the impact of falling household
sizes, and increasing household numbers, on
biodiversity changes should have high
Small households have adverse effects on
resource consumption both because they
are less energy-efficient in themselves and
news and views
30 JANUARY 2003
The threat of small households
Many studies have suggested that the increasing global human population
is having a negative effect on biodiversity. According to new work, another
threat comes from the rising number of households.
Average number of
people per household
1950 1970 1985 2000
World Less developed regions More developed regions
Figure 1Decline and fall in household sizes. Data for 1950 and 1970 are taken from ref. 8; data for
1985 and 2000 are from ref. 17.
because they often reflect an increase in the
number of households. If this increase could
be stabilized at roughly the same level as
population growth, the adverse effects might
also stabilize. But could this be achieved?
That depends on possible explanations for
why household sizes have fallen in the first
place. Some of these explanations are as
follows. First, all other factors remaining the
same, falling birth rates reduce population
size, but do not affect the number of house-
holds; hence, household size is reduced.
Second, increased material standards of
living have an effect. Extended households
are observed in countries in an early stage of
. When these countries attain
a higher standard of living, some institutions
— such as social-security systems — provide
the assurance against risks that were formerly
supplied by the extended household.
Third, social, economic and cultural
theories of demographic behaviour point to
a variety of reasons why individuals prefer
to live in small households
. These include
less adherence to strict norms; less religiosity
and increased individual freedom on ethical
issues; female education, which has led to
women having greater economic indepen-
dence and also facilitates divorce; more
assertiveness in favour of symmetrical gender
roles; the contribution of women to the
labour market; increased economic aspira-
tions; and residential autonomy. Fourth, pop-
ulation ageing reduces household size. This is
a direct consequence of two facts: increased
longevity leads to longer periods of time when
children do not live with their parents; and
the greater mortality of men, together with
the usual age difference between spouses,
results in many widows who live alone.
Smaller households, then, are the result
of processes that cannot be reversed (such as
modern contraception and liberalization
from norms) or that we value for a number of
reasons (such as women’s emancipation). So
policy interventions will have to focus on the
average household resource consumption,
in order to combat the adverse effects of
Nico Keilman is in the Department of Economics,
University of Oslo, Blindern, N-0317 Oslo, Norway.
1. Liu, J., Daily, G. C., Ehrlich, P. E. & Luck, G. W. Nature 421,
530–533 (2003); advance online publication, 12 January 2003
2. Cohen, J. How Many People Can the Earth Support?
(Norton, New York, 1995).
3. Boserup, E. Population and Technological Change
(Univ. Chicago Press, 1981).
4. Simon, J. The Ultimate Resource 2 (Princeton Univ. Press, 1996).
5. Population, Environment and Development: The Concise Report
(United Nations, New York, 2001). http://www.un.org/esa/
6. Ehrlich, P. & Holden, J. Science 171, 1212–1217 (1971).
7. Crenshaw, E., Ameen, A. & Christenson, M. Am. Soc. Rev. 62,
8. MacKellar, F. L., Lutz, W., Prinz, C. & Goujon, A. Pop. Dev. Rev.
21, 849–865 (1995).
9. Cramer, J. Demography 35, 45–65 (1998).
10.Jiang, L. Population and Sustainable Development in China
(Thela Thesis, Amsterdam, 1999).
11.O’Neill, B. & Chen, B. Pop. Dev. Rev. (Suppl.) 28, 53–88 (2002).
12.Bongaarts, J. Pop. Stud. 55, 263–279 (2001).
13.Goode, W. World Revolution and Family Patterns (Free Press,
New York, 1963).
14.Van de Kaa, D. Pop. Bull. 42, 1–59 (1987).
15.Lesthaeghe, R. in Gender and Family Change in Industrial
Countries (eds Mason, K. & Jensen, A.) 17–62 (Clarendon,
16.Verdon, M. Rethinking Households (Routledge, London, 1998).
17.United Nations Centre for Human Settlements (Habitat) Cities
in a Globalizing World (Earthscan, London, 2001).
undamental advances in colloid science
often depend on physical models,
which are made by dispersing carefully
tailored particles, less than a micrometre in
size, in pure aqueous or organic liquids.
Such dispersions can be characterized by
methods such as light scattering and confocal
microscopy, and the physical and chemical
interactions between the particles, responsi-
ble for intriguing phases such as colloidal
crystals (which behave like atomic solids), can
be precisely controlled. On page 513 of this
issue, Yethiraj and van Blaaderen
new model system that can be tuned with an
electric field to display phase transitions and
unexpected crystalline structures.
Colloidal crystals first attracted interest
in the 1960s. In studies of the light scattered
from dilute dispersions, a transition was
detected from a disordered fluid to an
ordered body-centred-cubic (b.c.c.) crystal
when the screened (or reduced) Coulomb
repulsions between the colloidal particles
extended to length scales greater than the
. In fact, this transition can be
controlled: adding a small amount of salt
decreases the range of the repulsive force,
because the salt dissociates into ions that
enhance the screening. As a result, the
volume fraction (or density) of particles
at the transition increases, and a denser,
face-centred-cubic (f.c.c.) crystal structure
is favoured. Adding even more salt leads to
‘hard-sphere’ transitions — as though the
particles were effectively hard spheres, with
no Coulomb repulsion. Then, entropy —
generally considered to be a measure of
disorder — favours the f.c.c. crystal, as
the number of configurations available to a
particle localized about a lattice site in the
f.c.c. crystal exceeds those accessible in a
disordered fluid or the b.c.c. crystal
Although hard-sphere behaviour of poly-
mer-based colloids could be achieved in
model systems, there was a drawback: those
colloids were opaque at even moderate densi-
ties, so little could be learned about their
structure from light scattering. More trans-
parent dispersions were sought, such as silica
spheres coated with short hydrocarbon
chains in a nonpolar solvent that eliminates
. In the 1980s, these organo-
philic silicas and the aqueous lattices sufficed
for many studies of fluid-to-crystal transi-
tions and other colloidal phenomena. But
small silicas could not easily be made highly
uniform in size and there can be extra, van der
Waals attractions between the larger ones, so
better colloidal hard spheres were sought.
Eventually a standard emerged: poly(methyl-
methacrylate) (PMMA) spheres coated with
a low-molecular-weight polymer
In a solvent that also contains soluble
polymer, neighbouring spheres are pushed
together by osmotic pressure due to expul-
sion of polymer chains from small gaps
between the particles. This attractive force
increases roughly linearly with polymer
concentration and can easily cause a dilute
gas-like dispersion to condense into a
colloidal fluid, and then into a solid f.c.c.
crystal. In reality, the hard spheres pass
through an intermediate, random hexa-
gonal close-packed (r.h.c.p.)
phase and only
slowly convert to the f.c.c. structure. For
larger colloids or smaller polymer chains,
the transition directly from ‘gas’ to f.c.c.
crystal is more favourable
Thus long-range attractions or repul-
sions yield condensed phases with low den-
sity and coordination number, such as dense
fluid or b.c.c. crystal phases. Short-range
repulsions and attractions produce denser
f.c.c. crystals with higher coordination
number. But crystals with lower coordina-
tion numbers than the b.c.c. phase or more
complex structures have not been achieved
with spheres of a single size. Yethiraj and van
confront this issue by devising
a model system in which the forces between
particles can be tuned, combining a soft
repulsion with a long-range, anisotropic
The authors laced PMMA spheres (with
radii between 1 and 2 mm) with fluorescent
dye and dispersed them in an organic
mixture whose refractive index and density
were chosen to aid confocal imaging of the
spheres. The solvent also preserves sufficient
dielectric contrast for an applied electric
field to induce strong dipole–dipole inter-
news and views
30 JANUARY 2003
Tunable colloidal crystals
William B. Russel
Microscopic particles dispersed in a solvent — a colloidal dispersion —
can be a useful model for phase transitions and crystal nucleation. A
colloid that can be ‘tuned’ using an electric field is a valuable new tool.