Hubble Sequence and Galaxy Morphology

The Hubble sequence is the periodic table of galaxies: a single diagram, drawn by Edwin Hubble between 1926 and 1936, onto which almost every large galaxy in the nearby universe can be placed by eye. It is a two-pronged tuning fork — a handle of smooth ellipticals (E0 through E7) that splits at the lenticular (S0) transition into two parallel branches of disk galaxies, normal spirals (Sa, Sb, Sc) and barred spirals (SBa, SBb, SBc), with irregulars beyond the tips. The ordering runs from featureless spheroids to structured, gas-rich, rotating disks, and the astonishing thing is how much physics rides along a purely visual axis. Move along the fork and four independent properties change together: the bulge-to-disk ratio falls, the spiral arms unwind and fragment, the cold-gas fraction and star-formation rate climb, and the colours shift from red to blue. Underneath sits the deepest distinction of all — dynamical support. Ellipticals are "hot," held up against gravity by the random motions of their stars (a velocity dispersion σ); spiral disks are "cold," held up by ordered rotation (a circular speed v_rot). The ratio v_rot/σ climbs from nearly zero on the elliptical handle to well above one at Sc. This article treats Hubble's picture as a quantitative instrument, not a filing cabinet. We replace the eye with three measurements that survive to the present day: the ellipticity index n = 10(1 − b/a) for spheroid shape, the Sérsic R^1/n family for surface-brightness structure, and the virial scaling relations — Faber–Jackson for ellipticals, Tully–Fisher for spirals — that turn kinematics into a distance-independent handle on mass. The one thing the sequence is not is a timeline: "early" and "late" are historical labels, and no galaxy slides from E to Sc as it ages.

Key takeaways

The Hubble tuning fork sorts galaxies by visual structure: ellipticals E0–E7, lenticulars S0, spirals Sa–Sc with barred variants SBa–SBc, and irregulars.
The elliptical class encodes only apparent flattening, n = 10(1 − b/a); it is a projected quantity, so an intrinsically flat E7 seen face-on masquerades as a round E0.
Surface brightness is quantified by the Sérsic law I(R) = I_e exp{−b_n[(R/R_e)^{1/n} − 1]}: n = 4 (de Vaucouleurs R^1/4, b_4 = 7.669) fits ellipticals and bulges; n = 1 (exponential, b_1 = 1.678) fits disks.
Along the spiral branches toward Sc, the bulge-to-disk ratio falls, arms loosen and clump, gas fraction and star formation rise, and colours blue.
The support ratio v_rot/σ rises from ≈0 (pressure-supported ellipticals) to ≫1 (rotation-supported spirals); the virial theorem then gives Faber–Jackson L ∝ σ⁴ and Tully–Fisher L ∝ v_rot⁴.
The sequence is descriptive, not evolutionary — "early/late" are relics of a wrong idea and imply no time axis.
Key idea: the Hubble sequence orders galaxies by two coupled trends — falling spheroid dominance and rising disk structure, gas content, and rotational support — from hot pressure-supported ellipticals through lenticulars to cold gas-rich spirals; it classifies present-day structure, not an evolutionary track.

The core insight

The Hubble sequence orders galaxies by two coupled trends — decreasing spheroid (bulge) dominance and increasing disk structure, gas content, and rotational support — from smooth pressure-supported ellipticals through lenticulars to gas-rich rotating spirals and irregulars; it is a classification of structure, not an evolutionary track.

Where this lives today

Hubble's fork is still the coordinate system of galaxy studies, but the eye has been replaced three times over. Profile-fitting codes such as GALFIT decompose a survey image into a Sérsic bulge and an exponential disk, returning n, R_e, and a bulge-to-total ratio for millions of galaxies from SDSS and now Euclid — the quantitative descendants of E-versus-Sc. Citizen science industrialised the eye itself: Galaxy Zoo enlisted hundreds of thousands of volunteers to classify a million SDSS galaxies, and their votes now train the neural networks that type entire surveys automatically. Kinematics has meanwhile split the elliptical handle in two — integral-field surveys (SAURON, ATLAS3D, MaNGA) measure a spatially resolved rotation parameter λ_R and find that most "ellipticals" are actually fast-rotating, disk-like systems, with only the most massive being genuinely pressure-supported slow rotators, a physically deeper dichotomy than E0–E7. The environmental axis Hubble hinted at is now the morphology–density relation (Dressler 1980): early types dominate cluster cores, late types the field, tracing how ram-pressure stripping and mergers quench and transform disks. And at the frontier, JWST is finding well-ordered disks and bars at redshifts z > 3 — far earlier than models predicted the Hubble sequence could assemble, turning the fork into a clock for galaxy formation after all, just not the one Hubble imagined.

The physical picture

Picture two knobs. The first shrinks the central bulge relative to the disk; the second unwinds the spiral arms and pours in gas, dust, and blue star-forming clumps. Ellipticals have no disk knob at all — all bulge, no arms. Move down the fork and you dial the disk up and the bulge down, until at Sc/SBc you reach a thin, gas-rich, weakly-bulged spinning disk. The barred and unbarred branches are the same disk family seen with and without a central stellar bar, a genuine dynamical instability of a self-gravitating disk that roughly two-thirds of disk galaxies (the Milky Way included) develop.

Figure

Hubble's tuning-fork diagram: ellipticals E0 to E7 form the handle where the class index is n = 10(1 − b/a), so E0 is round (b/a = 1) and E7 is the flattest observed (b/a ≈ 0.3); the fork splits at the lenticular S0 into normal spirals Sa Sb Sc and barred spirals SBa SBb SBc, with the bulge-to-disk ratio falling and gas content rising toward Sc

Think of dynamical support as the physics beneath the picture. In an elliptical the stars swarm on randomly oriented orbits — a pressure-supported, "hot" system whose depth is read from the line-of-sight velocity dispersion σ. In a spiral disk the same stars circle in an ordered, coplanar sheet — a rotation-supported, "cold" system read from the circular speed v_rot. Colour is the shorthand that ties it together: ellipticals populate the "red sequence" of old stars, late spirals and irregulars fill the "blue cloud" of ongoing star formation, and S0s bridge the two. What the eye grades as morphology, a spectrograph reads as star-formation history. The second figure shows the measurement that made this quantitative — the radial surface-brightness profile, the fingerprint that separates a spheroid from a disk without ever naming a Hubble type.

Figure

Surface-brightness profiles that quantify the sequence: a classical elliptical follows the de Vaucouleurs R^1/4 law μ − μe = 8.327((R/Re)^1/4 − 1) with Sérsic index n = 4 and b4 = 7.669, giving a sharp core and extended envelope, while a spiral disk follows the exponential μ = μ0 + 1.086 R/h — the magnitude axis is inverted so brighter is up

The core equations

Morphology becomes physics through four relations. The first turns the eye's grade of "roundness" into a number; the second and third replace the whole visual sequence with a two-parameter light profile; the fourth converts the invisible axis — how the galaxy holds itself up — into a measurable scaling law. Every symbol below is something a photometer or spectrograph reports directly. The elliptical class is a shape index. Hubble measured the apparent major and minor axes a and b on a photographic plate and quoted a single integer:

E_n:\qquad n = 10\left(1 - \frac{b}{a}\right) \qquad (1)

where: E_n — the elliptical subtype; a, b — apparent semi-major and semi-minor axes on the sky; n — rounded to an integer 0–7. Reading: E0 is a circle (b/a = 1); E7, the flattest elliptical ever catalogued, has b/a ≈ 0.3. Crucially n reports the projected shape only — it is a lower bound on the intrinsic flattening, not the flattening itself. The structural workhorse is the Sérsic profile, a one-parameter family of surface-brightness laws that absorbs the entire fork into the single index n:

I(R) = I_e\,\exp\!\left\{-b_n\left[\left(\frac{R}{R_e}\right)^{1/n} - 1\right]\right\} \qquad (2)

where: I(R) — surface brightness at projected radius R; R_e — the effective (half-light) radius, enclosing half the total light; I_e — surface brightness at R_e; n — Sérsic index; b_n — a constant, not free, fixed by the half-light definition to b_n ≈ 2n − 1/3. Reading: n = 4 recovers the de Vaucouleurs R^1/4 law of ellipticals and bulges (b_4 = 7.669); n = 1 is the exponential disk of late-type spirals (b_1 = 1.678). The single number n slides you along the sequence. Astronomers work in magnitudes, μ = −2.5 log₁₀ I, where the two anchor profiles become straight lines against the right radial variable — the form actually fitted to images and drawn in the figure above:

\mu(R) - \mu_e = 8.327\left[\left(\frac{R}{R_e}\right)^{1/4} - 1\right],\qquad \mu(R) = \mu_0 + 1.086\,\frac{R}{h} \qquad (3)

where: μ — surface brightness in mag/arcsec²; 8.327 = 2.5 log₁₀(e) × 7.669; h — the disk scale length; μ₀ — the extrapolated central disk brightness. Reading: the de Vaucouleurs law is a straight line versus R^1/4, the exponential disk a straight line versus R; the two slopes, 8.327 and 1.086, are the signatures a decomposition code separates in a real spiral. The unseen axis — pressure versus rotation — surfaces through the virial theorem as two flagship scaling laws:

L \propto \sigma^{4}\ \text{(Faber–Jackson, ellipticals)},\qquad L \propto v_{\rm rot}^{\,4}\ \text{(Tully–Fisher, spirals)} \qquad (4)

where: L — total luminosity; σ — central line-of-sight velocity dispersion (pressure support); v_rot — flat circular speed of the rotation curve (rotation support). Reading: both exponents follow from the virial theorem plus roughly constant mass-to-light ratio and surface brightness (Derivation C); the support ratio v_rot/σ rises from ≈0 on the elliptical handle to ≫1 at Sc, and its drift is what physically orders the fork.

Derivation AThe ellipticity index and the projection trap

The goal is to build the integer classifier n = 10(1 − b/a) from what a plate actually shows — the apparent axis ratio — and then to prove that this number can only bound a galaxy's true three-dimensional flattening, never fix it, because every elliptical is seen in projection from one unrepeatable direction. We define the index, project a model spheroid onto the sky as a function of viewing angle, and read off the limits, so that "E3" is understood as a statement about a shadow rather than about the object casting it.

Step 1. Take the apparent isophotes of an elliptical to be concentric ellipses of axis ratio q′ = b/a measured on the sky, and define the class by n = 10(1 − q′), rounded to an integer. The observed range 0 ≤ n ≤ 7 means 0.3 ≲ q′ ≤ 1; no relaxed elliptical flatter than q′ ≈ 0.3 is found. Why: the classifier must be computable from a single image, so it can depend only on the two apparent axes, nothing intrinsic.

Step 2. Model the true shape as an oblate spheroid with intrinsic axis ratio q = c/a < 1 (short symmetry axis c), viewed at inclination i, where i = 90° is edge-on and i = 0° is face-on down the symmetry axis. Projecting the ellipsoid onto the plane of the sky gives the apparent axis ratio q′² = cos²i + q² sin²i. Why: this is the exact map from the intrinsic shape (q) and the accident of orientation (i) to the only thing observable (q′).

Step 3. Read the limits. Edge-on (i = 90°): q′ = q, the galaxy shows its true flattening. Face-on (i = 0°): q′ = 1, the same galaxy looks perfectly round — an E0. So the measured n is a lower bound on intrinsic flattening; a genuine E7 can appear as any class down to E0. Only the statistics of q′ over a randomly oriented population recover the intrinsic shape distribution. Why: one line of sight cannot invert a three-parameter shape, so morphology from a single image is fundamentally degenerate — the origin of the pitfall below.

Go deeper · Derivation B · 2 minThe Sérsic law and why b_n is not free

Derivation BThe Sérsic law and why b_n is not free

The goal is to show that a single two-parameter family, I(R) = I_e exp{−b_n[(R/R_e)^{1/n} − 1]}, spans the whole photometric sequence — de Vaucouleurs ellipticals at n = 4, exponential disks at n = 1 — and to prove that the constant b_n is completely determined by insisting that R_e enclose exactly half the light. We integrate the profile in closed form, impose the half-light condition, and evaluate b_n at the two anchor indices, ending at the magnitude slope 8.327 that ties the algebra to the figure.

Step 1. Integrate the profile over the face of the galaxy: L = 2π ∫₀^∞ I(R) R dR. Substituting u = b_n (R/R_e)^{1/n} turns this into a gamma integral, giving the exact total luminosity L = 2π n I_e R_e² e^{b_n} b_n^{−2n} Γ(2n). Why: a closed form is what lets R_e, I_e, and n be fitted to data and compared between galaxies of different sizes and brightnesses.

Step 2. Demand that the light inside R_e be exactly L/2. The interior integral is the same gamma integral cut at u = b_n, so the condition becomes the transcendental equation γ(2n, b_n) = ½ Γ(2n), where γ is the lower incomplete gamma function. This has one solution for each n, with the accurate asymptotic form b_n ≈ 2n − 1/3 + 4/(405 n). Why: b_n is whatever makes R_e the half-light radius — it carries no independent physics, it is bookkeeping that gives R_e a universal meaning across all indices.

Step 3. Evaluate. For n = 4, γ(8, b) = ½ Γ(8) gives b_4 = 7.669 — the de Vaucouleurs R^1/4 law. For n = 1, γ(2, b) = ½ Γ(2) gives b_1 = 1.678, and since a pure exponential disk I = I₀ e^{−R/h} has its half-light radius at R_e = 1.678 h, the two match exactly. Converting the n = 4 case to magnitudes multiplies b_4 by 2.5 log₁₀(e) = 1.0857, giving μ − μ_e = 8.327[(R/R_e)^{1/4} − 1] — the straight-line-versus-R^1/4 slope of the plotted elliptical. Why: one family, one fitted index n, reproduces both ends of the fork and connects directly to the measurable magnitude profile.

Go deeper · Derivation C · 1 minPressure support, rotation support, and Faber–Jackson

Derivation CPressure support, rotation support, and Faber–Jackson

The goal is to make precise the deepest axis of the sequence — whether a galaxy stands against gravity by random motion or ordered rotation — and to derive the two flagship scaling laws, Faber–Jackson L ∝ σ⁴ for ellipticals and Tully–Fisher L ∝ v_rot⁴ for spirals, from a single argument: the virial theorem closed by two empirical regularities. The same three lines of algebra produce both laws, which is why deviations from them (the "tilt" of the Fundamental Plane) are read as slow drifts in those regularities rather than new physics.

Step 1. Apply the virial theorem, 2T + U = 0, to a self-gravitating stellar system. For a pressure-supported spheroid the kinetic term is set by the dispersion, giving M ≈ α σ² R_e / G with α a structure constant of order unity; for a rotation-supported disk the circular-orbit balance gives v_rot² ≈ G M / R. Why: this is the only step that contains gravity — it ties the mass to the one kinematic quantity each family offers, σ or v_rot.

Step 2. Add two observed regularities. First, old stellar populations give a nearly constant mass-to-light ratio, M ∝ L. Second, galaxies of a given family sit near a common effective surface brightness I_e, so L ∝ I_e R_e² ∝ R_e². Why: the virial theorem alone has three unknowns (M, R, kinematics); these two empirical closures reduce it to a single relation between L and the kinematic speed.

Step 3. Eliminate M and R_e. From M ∝ σ² R_e and M ∝ L: L ∝ σ² R_e. From L ∝ R_e²: R_e ∝ L^{1/2}. Substituting, L ∝ σ² L^{1/2}, hence L ∝ σ⁴ — Faber–Jackson. Repeating with v_rot in place of σ gives L ∝ v_rot⁴ — Tully–Fisher. The support ratio v_rot/σ therefore labels position on the fork, ≈0 for ellipticals and ≫1 for late spirals. Why: the two headline laws of galaxy scaling are one derivation seen from the hot and cold ends of the sequence, which is exactly the unity Hubble's picture hints at.

Limiting cases

Case	Result	Meaning
Round elliptical (E0)	b/a = 1, n = 0	apparent circle; may be intrinsically flat seen face-on
Flattest elliptical (E7)	b/a ≈ 0.3	empirical limit; flatter relaxed spheroids are not found (bar/bending instabilities)
Sérsic n = 1	exponential disk, b₁ = 1.678, R_e = 1.678 h	pure rotation-supported disk — bulgeless late-type (Sc/Sd)
Sérsic n = 4	de Vaucouleurs R^1/4, b₄ = 7.669	classical elliptical or a spiral's bulge
v_rot/σ ≫ 1	cold, rotationally supported	disk galaxy, late Hubble type
v_rot/σ ≪ 1	hot, pressure supported	elliptical, early Hubble type
Face-on projection (i → 0)	q′ → 1 for any intrinsic q	every spheroid can masquerade as E0

Common trapThe sequence is not a clock, and E-numbers are not shapes

Two traps hide in Hubble's vocabulary. First, "early-type" (E, S0) and "late-type" (Sc, Irr) are historical labels left over from a discarded belief that galaxies evolve along the fork from E to Sc as they age. They do not: a massive elliptical is typically older than a late spiral, and mergers can turn disks into spheroids — the opposite of the naming. Position on the fork reflects formation history, angular momentum, environment, and merger history, not a single time axis. Second, the ellipticity index E_n reports only the projected flattening on the sky, so a genuinely flattened E7 seen face-on is indistinguishable from a round E0 (Derivation A). Neither the "early/late" axis nor the E-number carries the physical content a beginner assumes it does.

Worked examples

1 · Classifying an elliptical, and the projection it hides (fully worked)

A relaxed elliptical shows apparent semi-axes a = 30″ and b = 21″. (i) Give its Hubble class. (ii) If it is intrinsically an oblate E7 (true axis ratio q = 0.3), at what inclination would it present the class you found?

Step 1. Apparent axis ratio: b/a = 21/30 = 0.70. The ellipticity index is n = 10(1 − 0.70) = 3.0, so the galaxy is classified E3. This is the only information a single image gives directly.

Step 2. Now suppose the true shape is much flatter, q = 0.30 (an intrinsic E7). The projection relation from Derivation A is q′² = cos²i + q² sin²i. Set q′ = 0.70 (so q′² = 0.49) and q² = 0.09: 0.49 = cos²i + 0.09(1 − cos²i) = 0.09 + 0.91 cos²i, giving cos²i = (0.49 − 0.09)/0.91 = 0.4396.

Step 3. Therefore cos i = 0.663, i.e. i ≈ 48° (measuring i from face-on, with 90° edge-on). A genuinely flattened E7 tilted about 48° from edge-on looks exactly like a round-ish E3. The lesson is quantitative, not just cautionary: the same galaxy spans E7 (edge-on) to E0 (face-on) purely through orientation, which is why intrinsic shapes are recovered only from the statistics of many galaxies.

2 · The concentration and total light of an R^1/4 elliptical (fully worked)

An elliptical is fitted by a de Vaucouleurs profile with effective radius R_e = 4 kpc and effective surface brightness μ_e = 21.5 mag/arcsec². (i) Find μ at R = 1 kpc and R = 16 kpc. (ii) Express its total luminosity in terms of I_e and R_e.

Step 1. Use μ(R) − μ_e = 8.327[(R/R_e)^{1/4} − 1] from equation (3). At R = 1 kpc, (1/4)^{1/4} = 0.707, so μ = 21.5 + 8.327(0.707 − 1) = 21.5 − 2.44 = 19.06 mag/arcsec². At R = 16 kpc, (16/4)^{1/4} = 4^{1/4} = 1.414, so μ = 21.5 + 8.327(0.414) = 21.5 + 3.45 = 24.95 mag/arcsec².

Step 2. Interpret the spread. Over a 16-fold range in radius the surface brightness drops by 24.95 − 19.06 = 5.89 mag, a factor 10^{5.89/2.5} ≈ 230 in intensity — the sharp bright core and faint extended envelope that make an R^1/4 spheroid unmistakable next to the gentle exponential fall of a disk.

Step 3. Total light from the Derivation B formula L = 2π n I_e R_e² e^{b_n} b_n^{−2n} Γ(2n). For n = 4, b_4 = 7.669, Γ(8) = 5040: the numerical prefactor is 2π·4·e^{7.669}·7.669^{−8}·5040 = 22.7, so L = 22.7 I_e R_e² = 7.22 π I_e R_e². This closed number is the standard aperture correction used to turn a fitted I_e and R_e into a galaxy's luminosity — the quantitative payoff of replacing the eye with a profile.

Physics of the portrait

The Hubble sequence is inseparable from the man who drew it: it was Hubble's second revolution, following hard on his proof in 1925 that the "spiral nebulae" were external galaxies at all.

Origins · 1 minEdwin Hubble and how this was discovered

Edwin Hubble

1889–1953 · extragalactic astronomy and the classification of galaxies

Edwin Powell Hubble came to astronomy late — he trained as a lawyer at Oxford on a Rhodes Scholarship and served in the First World War before joining the Mount Wilson Observatory, home to the 100-inch Hooker telescope, then the largest in the world. In 1924–1925 he found Cepheid variable stars in the Andromeda "nebula" and used the period–luminosity relation to prove it lay far outside the Milky Way, ending the "Great Debate" and establishing that the universe is filled with galaxies. With a universe of galaxies suddenly needing order, Hubble supplied it: his 1926 paper in the Astrophysical Journal laid out the classification of extragalactic nebulae into ellipticals, spirals, and irregulars, and his 1936 book The Realm of the Nebulae presented the tuning-fork diagram in the form still taught. He borrowed the "early/late" language from stellar spectral types and, wrongly, hinted it might be evolutionary — a caution he himself later voiced. Hubble is more famous for the velocity–distance law that bears his name (1929), but the morphological sequence is his other lasting instrument: nearly a century on, every galaxy survey still reports a Hubble type, and every quantitative refinement — de Vaucouleurs' three-dimensional scheme, Sérsic's profile, Galaxy Zoo's million eyes, today's neural classifiers — is a footnote to the fork he sketched.

History

1926

— Hubble classifies extragalactic nebulae into ellipticals, spirals, and irregulars

1936

— Hubble's The Realm of the Nebulae presents the tuning-fork diagram

1948

— de Vaucouleurs fits the R^1/4 surface-brightness law to elliptical galaxies

1961

— Sandage's Hubble Atlas of Galaxies codifies the visual sequence

1968

— Sérsic generalizes the profile to R^1/n, unifying bulges and disks

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 25 (Galaxies) · Binney & Tremaine, Galactic Dynamics, 2nd ed., Ch. 1 and §4.3 · Mo, van den Bosch & White, Galaxy Formation and Evolution, Ch. 2 and Ch. 11 · Sparke & Gallagher, Galaxies in the Universe, 2nd ed., Ch. 5–6 · Ryden, Introduction to Cosmology, 2nd ed., Ch. 1. Key numerical checks: $b_4 = 7.669$ and $b_1 = 1.678$ follow from solving $\gamma(2n,\,b_n) = \tfrac{1}{2}\Gamma(2n)$ ; the magnitude slope $8.327 = 2.5\log_{10}(e)\times 7.669$ and the disk slope $1.086$ follow directly; an exponential disk has $R_e = 1.678\,h$ ; the total light of an $n=4$ profile is $L = 22.7\,I_e R_e^2 = 7.22\pi\,I_e R_e^2$ ; and the projection relation $q^{\prime 2} = \cos^2 i + q^2\sin^2 i$ yields $i \approx 48°$ for the worked example.

Prerequisites. Structure of the Milky Way Galaxy · Stellar Populations I, II, and III. Leads to. Elliptical Galaxies and the de Vaucouleurs Profile · Spiral Galaxies and Density Wave Theory · Galactic Bulge, Bar, and Central Regions · Interstellar Medium of Galaxies.

Ask the tutor: Ask me to fit a bulge-plus-disk Sérsic decomposition to a real galaxy, or to walk through why v_rot/σ — not the eye — is the sequence's true physical axis.

Is the Hubble sequence an evolutionary sequence — do galaxies move from E to Sc?

No, and this is the single most common misconception. The "early/late" language borrowed from stellar classification wrongly suggested a time axis, but galaxies do not slide along the fork as they age. If anything the arrow points the other way: massive ellipticals host the oldest stellar populations, and the leading channel for making an elliptical is a merger of two disk galaxies — a late-to-early transformation. Position on the sequence encodes formation history, angular momentum, environment, and merger history, all frozen into present-day structure. The sequence is a snapshot classification of what galaxies are, not a filmstrip of what they become.

Why does the sequence stop at E7 — why are there no E8 or E9 ellipticals?

Because relaxed, pressure-supported spheroids flatter than about b/a ≈ 0.3 are dynamically unstable and simply do not form as ellipticals. A stellar system flattened by rotation would have become a disk; one flattened by anisotropic random motions runs into bending and bar-like instabilities that thicken it back up before it reaches E8. So E7 is an empirical wall set by stellar dynamics, not a limit of the classification scheme. Anything you might be tempted to call "E8" is really an edge-on disk or S0 whose thin light distribution betrays rotational, not pressure, support.

What separates a lenticular (S0) from an elliptical?

An elliptical is a single pressure-supported spheroid with no disk; a lenticular has a genuine, often rotating, stellar disk plus a bulge, but no spiral arms and little cold gas or star formation. Structurally, an S0 needs a two-component fit — an R^1/4-like bulge and an exponential disk — while a pure elliptical is fitted by one Sérsic profile. Physically the S0 is best read as a "used-up" spiral: a disk galaxy that lost or exhausted its gas (through stripping in clusters, or internal quenching) and stopped forming the young blue stars that trace arms. That is why S0s cluster at the fork's junction and bridge the red sequence and blue cloud in colour.

Are barred spirals a fundamentally different class from normal spirals?

No — SB and S are the same disk family wearing or not wearing a bar. A stellar bar is a self-consistent, rotating standing-wave pattern that a cold, self-gravitating disk spontaneously develops when it is dynamically responsive; roughly two-thirds of disk galaxies, including the Milky Way, host one. Bars come and go over a galaxy's life and channel gas inward, feeding central star formation and bulge growth, so "barred" is a phase and a driver of evolution rather than a separate species. Hubble drew the two branches in parallel precisely because everything else — bulge fraction, arm winding, gas content — varies the same way along each.

What is the Sérsic index, and how does it map onto Hubble type?

The Sérsic index n is the single shape parameter in I(R) = I_e exp{−b_n[(R/R_e)^{1/n} − 1]}: it controls how centrally concentrated the light is. Low n means a shallow core and a sharp edge (n = 1 is the exponential disk of late-type spirals); high n means a bright cusp and a far-reaching envelope (n = 4 is the de Vaucouleurs law of ellipticals and bulges, with n = 6–8 for the most massive cluster giants). Because concentration tracks bulge dominance, n correlates well with Hubble type — early types have high n, late types low n — which is why automated surveys often quote n as a continuous, quantitative stand-in for the visual class.

How do modern surveys classify galaxies — by eye or by machine?

Both, in a pipeline. The eye set the ground truth: the Galaxy Zoo project had citizen scientists visually classify a million SDSS galaxies, producing a labelled catalogue no single expert could match in volume. Those human labels then trained convolutional neural networks that now morphologically type entire surveys automatically and consistently. In parallel, profile-fitting codes report objective structural numbers — Sérsic n, effective radius, bulge-to-total ratio, asymmetry, concentration — and integral-field spectrographs add kinematic classes (fast versus slow rotator). Modern "morphology" is therefore a fusion of visual labels, fitted structure, and resolved kinematics, all still anchored to the Hubble types they refine.

Where does the Milky Way sit on the Hubble sequence?

The Milky Way is a barred spiral, classified roughly SBbc — an intermediate-to-late-type disk with a modest bulge, a central bar (established from infrared star counts and the kinematics of bulge stars), and a bulge-to-total light ratio of only about 0.1–0.2. That places us between Sb and Sc on the barred branch, with Andromeda (M31) a somewhat earlier, more bulge-dominated Sb. Living inside the disk makes our own type hard to measure — we infer it the way we infer another galaxy's, from star counts, the rotation curve, and gas content, never from a face-on portrait.