2. Evolution under Artificial Selection

Oil Content in Corn

Artificial selection has been carried out on a variety of traits in a number of organisms.

Although some examples only a biologist could love, other examples have had an important impact on agriculture, including selection on

• birth weight, growth rate, and milk production in cows
• egg production in chickens
• back-fat in pigs
• grain yield in wheat

One of the longest running studies where evolutionary change has been documented began in corn in 1896 at the University of Illinois.

Two lines of corn were artificially selected. In one line, those plants with a high oil content were used as seed in the next generation. In the second line, plants with a low oil content were used as seed.

[Dudley and Lampert (1992)]

(Why has the lower line tapered off?)

2. Evolution under Artificial Selection

Body Weight in Mice

More complicated traits also respond to artificial selection.

For example, the weight of mice at six weeks of age was selected, again in two separate lines (for heavier and for lighter mice).

[Roberts (1966)]

(The dashed lines refer to a subset of the two lines in which artificial selection was reversed.)

(The dotted lines refer to a subset of the two lines in which artificial selection was stopped.)

2. Evolution under Artificial Selection

Thorax Length in Drosophila

Thorax length in adult Drosophila melanogaster shows a similar response to selection.

Here two separate lines were selected: one for longer flies and one for shorter flies.

[Robertson 1955]

(The dashed lines refer to a subset of the two lines in which artificial selection was reversed.)

(The dotted lines refer to a subset of the two lines in which artificial selection was stopped.)

2. Evolution under Artificial Selection

Abdominal Bristle Number

The number of bristles is a trait that is fairly easy to score in Drosophila melanogaster.

Yoo (1980) selected for an increased number of bristles in six replicate lines.

The number of bristles changed from around ten to around forty in 90 generations!

After about 90 generations, selection was stopped (*).

(Why would the number of bristles decrease after artificial selection was stopped?)

As long as the initial population is genetically variable, artificial selection is almost always successful and the trait under selection changes over time.

Even starting with a genetically homogeneous population, artificial selection still works, but it takes longer since selection is limited by a lack of variation and acts only on new mutations that occur.

For example, Mackay et al (1994) selected on abdominal bristle number in a highly inbred line of Drosophila (=extremely low in genetic variability).

Nevertheless, over 120 generations the high and low lines differed by 12 bristles on average!

OK - So a trait changes over time under selection, but could that ever lead to two different species?

2. Evolution under Artificial Selection

Habitat Selection in Drosophila

Rice and Salt (1988, 1990) designed an experiment to test the hypothesis that selection could drive speciation among sympatric populations.

Definitions: from Bush and Howard (1986)

Species: "A group of populations whose evolutionary pathway is distinct and independent from that of other groups...achieved by the group's reproductive isolation from other groups."

Sympatry: "Two populations are sympatric if individuals of each are physically capable of encountering one another...with moderately high frequency."

Allopatry: "Allopatric populations are separated by uninhabitated space...across which migration (movement) occurs at a very low frequency."

Rice and Salt constructed an ingenious maze within which larval flies were placed.

Within the maze, flies could choose

• lightness or darkness by moving left or right [selection for phototaxis]
• up or down [selection for geotaxis]
• acetaldehyde (white vial) or ethanol (black vial) [selection for chemotaxis]

In addition, flies were collected from the eight "habitats" during three time periods: early (E), middle (M), and late (L) [selection for development time].

Flies were allowed to mate within the maze (females tend to mate only when they have located food).

For the control lines, all flies within the habitats were mixed and 120 females chosen.

For the selected lines, 60 females flies were drawn from habitat 5E (dark, up, acetaldehde) and 60 from habitat 4L (light, down, ethanol).

Larvae from the experimental females were mixed and placed together in the maze to start the next generation. Controls were run through the maze separately.

(Offspring of mothers collected from 5E and half of the controls were raised on a chemical that turned their eyes brown.)

Over the 35 generations of the experiment, habitat specialization evolved in the selected lines:

• the offspring of females collected from habitat 5E (solid squares) were more likely to return to 5E
• the offspring of females collected from 4L (empty squares) were more likely to return to 4L.

No habitat specialization evolved in the control flies.

Since females tend to mate near the food vials, gene flow between 5E and 4L flies had virtually ceased by the end of the experiment.

The first step of speciation had occurred!

SOURCES:

• Rice and Salt (1988) Speciation via disruptive selection on habitat preference: Experimental evidence. American Naturalist 131: 911-917.
• Rice and Salt (1990) The evolution of reproductive isolation as a correlated character under sympatric conditions: Experimental evidence. Evolution 44: 1140-1152.