ASTR 101 ยท Lecture 12

Galaxies: Types, Interactions, and Evolution

Evidence, model, and quantitative reasoning

Learning Goals

  1. Explain why this lecture's central phenomenon matters: galaxies are not isolated static islands.
  2. Use the key model idea that morphology is a useful description but not a complete evolutionary sequence.
  3. Apply the lecture's quantitative tool with units, assumptions, and a physical interpretation.
  4. Correct the misconception: The Hubble tuning fork is a classification diagram, not a simple evolutionary path from one type to another.

Reading anchor: OpenStax Astronomy 2e, Chapters 26.1-26.5 and 27.1-27.4 on galaxy types, distances, active galaxies, and galaxy evolution.

Why This Question Matters

The Curtis-Shapley debate and Hubble's distance measurements moved galaxies from 'spiral nebulae' to island universes. Morphology became a clue, not the whole story.

Opening Phenomenon

Galaxies are not isolated static islands. Their shapes, colors, star formation, and activity carry evidence of environment and history.

First question: what is directly observed, and what part is an explanation built from a model?

Terms for This Lecture

spiral galaxy

Definition: A galaxy with a rotating stellar disk, gas and dust, and often spiral arms surrounding a central bulge.

Why it matters here: Its structure connects angular momentum, ongoing star formation, and environmental history.

Relationships: Spirals tend to contain colder gas and bluer young populations than many ellipticals.

Boundary or common confusion: Morphology alone does not specify mass, age, or a fixed evolutionary stage.

Evidence We Need to Explain

Physical or Geometric Model

Quantitative Tool

\[ z = \frac{\lambda_{\mathrm{obs}} - \lambda_0}{\lambda_0} \]

At low redshift, redshift can approximate recession speed; at larger redshift, cosmology matters.

Worked Example

  1. A spectral line with rest wavelength 500.7 nm observed at 550.8 nm has z about 0.10.
  2. At low redshift, this corresponds approximately to recession speed 0.10c.
  3. For larger redshift, cosmological interpretation requires a more careful model.

The answer should explain when the low-redshift approximation is adequate and when it is not.

Visual Reasoning

Use morphology, color, spectra, and environment as evidence for galaxy history, not as labels alone.

  • Which features indicate interaction?
  • What evidence traces recent star formation?
  • Why is morphology not a full evolutionary history?
ESA/Hubble image of the interacting Antennae Galaxies.
Tidal distortion and compact blue star clusters show that galaxy morphology can record interaction history. Credit: NASA, ESA, and the Hubble Heritage Team (STScI/AURA)-ESA/Hubble Collaboration; acknowledgement B. Whitmore and James Long; CC BY 4.0.

Common Misconception

The Hubble tuning fork is a classification diagram, not a simple evolutionary path from one type to another.

Correct it by treating morphology as classification, not a required evolutionary sequence.

Active Learning Segment

Students classify galaxy images, justify criteria, and identify at least one ambiguous or interaction-disturbed case.

Write a prediction first, compare reasoning with a neighbor, then revise your answer in one sentence.

Lab or Observing Connection

Lab 08 applies documented morphology criteria and simple redshift interpretation. Your classifications should cite morphology criteria and note ambiguous or interacting systems.

Synthesis

Galaxies connect local astrophysics to cosmic history: stars, gas, black holes, dark matter, and environment all matter.

References and Visual Credits